CN108052507A - A kind of city management information the analysis of public opinion system and method - Google Patents
A kind of city management information the analysis of public opinion system and method Download PDFInfo
- Publication number
- CN108052507A CN108052507A CN201711473257.9A CN201711473257A CN108052507A CN 108052507 A CN108052507 A CN 108052507A CN 201711473257 A CN201711473257 A CN 201711473257A CN 108052507 A CN108052507 A CN 108052507A
- Authority
- CN
- China
- Prior art keywords
- emotion
- participle
- analysis
- word
- negative
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/3331—Query processing
- G06F16/3332—Query translation
- G06F16/3335—Syntactic pre-processing, e.g. stopword elimination, stemming
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/3331—Query processing
- G06F16/334—Query execution
- G06F16/3344—Query execution using natural language analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/335—Filtering based on additional data, e.g. user or group profiles
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- General Health & Medical Sciences (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present invention relates to a kind of city management information the analysis of public opinion system and methods, and database is segmented including structure；Realize the acquisition of text data；Emotion pre-processes after Chinese word segmentation and participle；It is filtered after participle；Hot spot and sentiment analysis；Result is subjected to visualization processing and preserves and exports with Excel forms.The beneficial effects of the invention are as follows：Public sentiment is analyzed using segmentation methods and emotion vector algorithm, excavate the hot spot of city management information public sentiment and emotion direction, the specific aim of work is further strengthened for the reaction force of city management construction by public sentiment, so as to improve the efficiency of municipal administration's work, yield, strengthen city management construction.The present invention proposes a kind of public sentiment sentiment analysis algorithm based on emotion vector, and effectively accurately emotion prediction can be carried out to public sentiment, so as to provide support for public sentiment hot statistics and public sentiment direction control.
Description
Technical field
The present invention relates to the analysis of public opinion system, more specifically, it be related to a kind of city management information the analysis of public opinion system and
Method.
Background technology
In recent years, with the fast development of internet, the network public opinion information flow-rate in China is always in the state to accelerate
Gesture, and acquisition of information and intercommunion platform are also being on the increase.Internet is promoting information interchange and while social progress,
Many problems and challenge are brought to city management, are mainly shown as the uncontrollability of social speech, explosion type public opinion is brought negative
Face social influence etc..Whenever having negative public opinion to break out, immeasurable negative social effect can be brought.
There is long history in China on the construction of public sentiment thought and system, but the theoretically really research to public sentiment
It starts from 2003, the research of network public-opinion is started from 2005.Because public sentiment research is a new social science and natural section
The research field intersected is learned, the personnel and organizations studied at home this are relatively fewer, and the depth of investigation also waits to strengthen.But
Some valuable achievements in research occurred in recent years, to understanding and studying the very enlightening meaning of network public-opinion.Public opinion can not be complete
It avoids entirely, artificial public opinion monitoring, can not be in negative carriage because of many restrictions such as its huge human cost and reaction time is slow
Public opinion management and control is carried out by the first time of outburst, effective public opinion information analysis is taken and identifies its affective characteristics, and to the whole network
Information carries out System and temperature analysis is the primary study content of municipal administration's the analysis of public opinion.The analysis system of municipal administration's information public sentiment
It is exactly a kind of effective non-engineering measure.
Patent 201610047697.7 " a kind of internet public feelings analysis method " proposes a kind of internet public feelings analysis side
Method, the internet public feelings analysis method include：First against selected acquisition event, microblogging source text is divided, removal with
The unrelated division item of mood；Then counted using statistical and analytical tool, obtain an input of mood disaggregated model；Finally
Correlation word, expression, symbol that mood can be expressed in content of microblog are modeled with sorting algorithm for input, provide synthesis
Affection index is evaluated, and is obtained mood classification, and is carried out public sentiment monitoring and mood trend analysis.The invention is to word, table in microblogging
Feelings and symbol etc. carry out mood modeling, are calculated by moos index, the reaction situation of focus incident in microblogging can be carried out automatic
Classification and effective monitoring.Patent 201410073473.4 " the analysis of public opinion method and system " proposes a kind of the analysis of public opinion method, bag
Include following steps：It is searched for according to searching request and reads web page files；Public feelings information is extracted from web page files；To public feelings information
Classify；It is further analyzed to obtain the public sentiment letter in each classification results to the public feelings information in each classification results
Cease corresponding origin, public opinion emotional color, network disperse state, development trend, regional information and age segment information；According to carriage
Whether the further classification results of feelings information and default preservation of evidence rule judgment carry out the preservation of evidence to public feelings information.This
A little method and systems only realize and the analysis of public opinion and qualitative are carried out to particular webpage or text, it is impossible to realize to city management information
Intelligent the analysis of public opinion.
The content of the invention
The purpose of the present invention is overcome in the prior art to functions such as city management information, sentiment analysis and hot word statistics
Deficiency provides a kind of city management information the analysis of public opinion system.
This city management information the analysis of public opinion system, includes the following steps：
Step 1: structure participle database：It is deposited using based on the database of Oracle to having natural language participle
It stores up and provides database for algorithm calculating and support；
Step 2: realize the acquisition of text data：Using the project management system based on Maven, text is carried out on foreground
The typing of data, with Ajax by data storage with Json in url carries out with the interaction on backstage so that server can obtain need
The text message to be analyzed；
Step 3: emotion pre-processes after Chinese word segmentation and participle：It is carried out by dismembering an ox as skillfully as a butcher algorithm to having text message
Basic word segmentation processing will segment deposit participle database and put on index and the meter of emotion value is carried out at the same time during participle
It calculates, according to set several feature vectors, these feature vectors are broadly divided into positive emotion qualifier and negative emotion qualifier,
It calculates and analyzes to be indexed the emotion value of comment or model further according to the sentiment analysis of participle；
Step 4: it is filtered after participle：Noise vocabulary, filtering can be also referred to as in one model there are many useless vocabulary
Work mainly by specific algorithm, is filtered by set keyword feature vector or keyword storehouse, after then filtering
Result be inserted into hot spot dictionary；
Step 5: hot spot and sentiment analysis：The hot spot dictionary that is calculated according to step 4 and each segment corresponding feelings
Sense vector carries out the calculating of emotion value and hot statistics to all analysis data, draws the corresponding emotion value of each sentence and participle
Temperature sequence, and analysis result is subjected to visualization processing；
Step 6: result is subjected to visualization processing and preserves and exports with Excel forms.
The step 1 specifically includes：Database based on Oracle table design, mainly include public sentiment table, participle table,
Table is segmented after keyword table, filtering, the field that the public sentiment table includes has number, content, time, source, the feelings of corresponding informance
Feel assay value, source address, the field that the participle table includes has participle number, participle content, the corresponding information of participle
Source number, part of speech, affection data, source-information content, the field that the keyword table includes have keyword number and keyword content,
The field that participle table includes after the filtering have segment content after participle number, filtering after filtering, the corresponding information of participle is come
Source number, participle frequency.
The step two specifically includes：Model content and number are inputted on foreground by Maven systems, and pass through Json
Data interchange format and url are interacted with server end, and content of text is passed to server and according to the data of public sentiment table
Form is preserved.
The step three specifically includes：
Using dismembering an ox as skillfully as a butcher, algorithm is segmented：
Character string is passed to algorithm of dismembering an ox as skillfully as a butcher to segment, participle is included with quotation marks after participle, participle is separated with space,
Participle form is rewritten, then removes the mark of word segmentation, leaves behind participle and space, this result is passed into principal function, principal function
The space number after participle is calculated to determine participle quantity, removes space, the character string after participle is formally switched to segment to be single
The array of position.
Emotion counts：
Established keyword table and participle table are read, imports positive emotion vector array, negative emotion vector array, emotion
Degree vector array and negative emotion vector array, are circulated in for and judge participle vector one by one, and emotion vector is used
Boolean variables export, positive emotion true, negative emotion false.
Emotion modification statistics：
The statistical analysis of emotion degree will be depending on emotion degree vector, and different emotion degree is compared with different numbers
According to, if centre word be negative word, by binary emotion be set to -1. weak first yuan for negative, then be set to -1, be multiplied with second yuan no
Fixed is negatively front, if there is qualifier before centre word, is multiplied by corresponding weight value to represent emotion degree, is first calculating last binary just
Face degree, then whether determine addition of vectors, draw emotion value result.
Emotion statistical analysis is carried out by triple method, it is 0 to set ternary each several part and total value initial value, remembers triple
Respectively shaping variable res1, res2, res3, note total value is shaping variable res, if centre word is negation words, sets the
Binary variable res2 is -1, if first yuan is divided into following three kinds of situations for negative：
If 1) first yuan be negative, res1 is set for -1, total value is made to be equal to res2 and is multiplied by res1, negative is negatively
For certainly；
If 2) there is qualifier before negation words, res1 is assigned a value of corresponding weight value, and res is made to be multiplied by res2 equal to res1, to say
Bright negative degree returns to total value res；
If 3) only second yuan of centre word, total value res is equal to the numerical value of centre word res2, returns to total value res；
When centre word res2 is positive emotion, statistical method is similar with above-mentioned negative statistical analysis.
When centre word is negative word, it is divided into following situation：
1) as only centre word res2, total value res is made to be equal to res2, returns to total value res0；
2) when there are total value res during qualifier res3, is made to be multiplied by res3, return total value res0 equal to res2；
3) when there are during qualifier, make res1 that res be made to be multiplied by res3 equal to res2 and add res1, returned equal to corresponding weights
Total value res0；
When centre word for conservative negative when, first calculate the positive degree of last binary, then whether determine addition of vectors, draw finally
As a result.If qualifier will negate vectorial to negate extremely vector and negative multiplication of vectors, its negative degree is drawn.
The step four specifically includes：
The word that emotion value is 0 is extracted in the participle table drawn by step 3, it is then basic in keyword storehouse
Information comparison filters out some noise words, i.e., nonsensical vocabulary, remaining word is mostly ranking, mainly including place name, day
Phase and name etc., these words can just become hot spot vocabulary, the result after filtering are put into hot spot dictionary., statistics is most
Subsection is public sentiment member, and so-called public sentiment member is exactly that vocabulary, source are bound together to the unit to be formed insertion hot spot filtering
In dictionary, in statistics, if the situation that vocabulary source and vocabulary are identical with having vocabulary in database, is not counted in system
Meter only adds up numerical value in the word frequency for having vocabulary.
The step five specifically includes：
Public sentiment sentiment analysis result chart：
There is provided the data of the sentiment analysis statistics of multi-angle, column map generalization and form in emotion comprehensive statistics module
Output function is compared, corresponding assigned variable is passed to by proxy according to foreground selection, backstage Json is obtained and is shown on foreground,
Positive Negative sentiments vector can be shown on combobox if selecting by vector analysis is tended to, if selecting according to source analysis,
It can show the site information in various sources.
Public sentiment hot counts：
What public sentiment hot statistics and analysis carried out on the basis of step 3 and step 4, the vocabulary in table is segmented to filtering
It extracts, is compared with the information in public sentiment table and carries out further statistical analysis again.
Focus statistics is based on word frequency, using public sentiment source as the unit of statistical analysis member, by centre word junction to backstage, is carried
Go out the hot spot vocabulary beyond the vocabulary of center, form a unduplicated storage array, source ratio then is carried out to the information of extraction
Right, the center vocabulary hot spot coefficient from same public feelings information source is arranged to 1, and hot spot coefficient is multiplied by the word frequency of corresponding centre word
That is the hot spot degree of the centre word；If vocabulary to be analyzed is not from same public sentiment source with centre word, word is analysed to
Site information is compared with public feelings information storehouse, and the degree of correlation that station external information and centre word are considered to be if inconsistent is
Zero, interior information of standing is considered as if consistent, closes on algorithm using URL to carry out its calculating with the related coefficient of centre word,
It is exactly its hot spot degree that related coefficient finally is multiplied by its word frequency, and analysis result is exported in units of word.
The step six specifically includes：
It is realized by key-course ExcelBuild classes and the ExcelUtil classes on backstage and exports data with Excel forms.
The beneficial effects of the invention are as follows:The present invention proposes a kind of city management information the analysis of public opinion system, system profit
Public sentiment is analyzed with segmentation methods and emotion vector algorithm, excavates hot spot and the emotion side of city management information public sentiment
To further strengthening the specific aim of work for the reaction force of city management construction by public sentiment, so as to improve municipal administration's work
The efficiency of work, yield strengthen city management construction.The present invention proposes a kind of public sentiment sentiment analysis based on emotion vector and calculates
Method can carry out public sentiment effectively accurately emotion prediction, so as to provide support for public sentiment hot statistics and public sentiment direction control.System
System can effectively store a large amount of participle data and keyword data and the rapid growth for tackling data, using number using Orace databases
The inherent value of historical data is excavated according to visualization technique, information guiding is provided for the analysis of public opinion and management and control work.
Description of the drawings
Fig. 1 is city management information the analysis of public opinion system function frame diagram proposed by the present invention；
Fig. 2 is city management information the analysis of public opinion system business process figure proposed by the present invention；
Fig. 3 is the public sentiment table structure chart that the present invention describes；
Fig. 4 is the participle table structure chart that the present invention describes；
Fig. 5 is the keyword table structure chart that the present invention describes；
Fig. 6 is Vocabulary structure figure after the filtering that the present invention describes；
Fig. 7 is the addition Operation interface diagram that the present invention realizes；
Fig. 8 is the public sentiment emotion statistic analysis result figure that the present invention realizes；
Fig. 9 is the vocabulary filtering and hot spot word frequency statistics result figure that the present invention realizes；
Figure 10 is the public sentiment sentiment analysis result block diagram that the present invention realizes；
Figure 11 is the public sentiment sentiment analysis result tabular drawing that the present invention realizes；
Figure 12 is the public feelings information hot spot comprehensive analysis block diagram that the present invention realizes；
Figure 13 is Excel datagrams derived from the public feelings information analysis result of the present invention.
Specific embodiment
The present invention is described further with reference to embodiment.The explanation of following embodiments is only intended to help to understand this
Invention.It should be pointed out that for those skilled in the art, without departing from the principle of the present invention, also
Can to the present invention some improvement and modification can also be carried out, these improvement and modification also fall into the protection domain of the claims in the present invention
It is interior.
The overall structure of the system as shown in Figure 1, operation flow as shown in Fig. 2, specific implementation step is as follows：
Step 1: structure participle database
By to text message segment data research it can be found that text message participle mainly include emotion segment,
Information segments and noise participle, and the public feelings information table of structure as shown in Figure 3 is established to storage source text envelope according to system structure
Breath, establishes the participle table of structure as shown in Figure 4 to store participle information, establishes the keyword table of structure as shown in Figure 5, establishes such as
Table is segmented after the filtering of structure shown in Fig. 6.
Step 2: realize the acquisition of text data
System employs the project management plug-in unit based on Maven, after addition operation as shown in Figure 7 is carried out on foreground, leads to
It crosses Ajax to interact data storage with backstage with the url set in Json, by interpolation data according to public feelings information table
Structure is stored in server, then is thrown to by the Json data formats on backstage and foreground and show latest result.
Step 3: Chinese word segmentation and emotion pretreatment
After the completion of data import, text message is passed to algorithm of dismembering an ox as skillfully as a butcher and is segmented, included point with quotation marks after participle
Word separates participle with space, and participle form is rewritten, then removes the mark of word segmentation, leaves behind participle and space, this result is passed
Principal function is passed, principal function calculates the space number after segmenting to determine participle quantity, removes space, by the character string after participle just
Formula switchs to the array in units of participle and is stored in the server according to participle table structure.
Established keyword table and participle table are read, imports positive emotion vector array, negative emotion vector array, emotion
Degree vector array and negative emotion vector array, the emotion vector array judged one by one described in participle table is circulated in for, will
Emotion vector is exported with boolean variables, positive emotion true, negative emotion false.
The statistical analysis of emotion degree is depending on emotion degree vector, and different emotion degree is compared with different numbers
According to, if centre word be negative word, by binary emotion be set to -1. weak first yuan for negative, then be set to -1, be multiplied with second yuan no
It is fixed negative to be positive, if there is qualifier before centre word, corresponding weight value is multiplied by represent emotion degree, calculates the front of last binary
Degree, then whether determine addition of vectors, draw emotion value result.
The participle information for carrying out participle and emotion pretreatment is as shown in Figure 8.
Step 4: it is filtered after participle
The word that emotion value is 0 is extracted in the participle table drawn according to step 3, it is then basic in keyword storehouse
Information comparison filters out noise word, and remaining word mainly includes place name, date and name etc., the result after filtering is put into heat
In point dictionary.When counting word frequency, if the source of vocabulary to be counted and vocabulary are identical with having vocabulary in database, no
Statistics is included in, only numerical value adds up in the word frequency for having vocabulary, the result after filtering is then inserted into hot spot dictionary.
Filter result is as shown in Figure 9 after participle.
Step 5: hot spot and sentiment analysis
(1) public sentiment hot counts
What public sentiment hot statistics and analysis carried out on the basis of step 3 and step 4, the vocabulary in table is segmented to filtering
It extracts, is compared with the information in public sentiment table and carries out further statistical analysis again.
By centre word junction to backstage, propose the hot spot vocabulary beyond the vocabulary of center, form a unduplicated storage number
Then group carries out source comparison to the information of extraction, the center vocabulary hot spot coefficient from same public feelings information source is arranged to
1, hot spot coefficient is multiplied by the hot spot degree of the i.e. centre word of word frequency of corresponding centre word；If vocabulary to be analyzed is not with centre word
From same public sentiment source, then the site information for being analysed to word is compared with public feelings information storehouse, is considered as if inconsistent
It stands external information, the degree of correlation with centre word is zero, is considered as interior information of standing if consistent, closes on algorithm using URL to carry out
Its calculating with the related coefficient of centre word, it is exactly its hot spot degree that related coefficient finally is multiplied by its word frequency, and in units of word
Export analysis result.
(2) public sentiment sentiment analysis result chart
The hot spot dictionary that is calculated according to step 4 and each segment corresponding emotion vector to all analysis data into
Market inductance value calculates and hot statistics, draws the temperature sequence of the corresponding emotion value of each sentence and participle, and analysis is tied
Fruit carries out visualization processing.
There is provided the data of the sentiment analysis statistics of multi-angle, column map generalization and form in emotion comprehensive statistics module
Output function is compared, corresponding assigned variable is passed to by proxy according to foreground selection, backstage Json is obtained and is shown on foreground,
Positive Negative sentiments vector can be shown on combobox if selecting by vector analysis is tended to, if selecting according to source analysis,
It can show the site information in various sources.
Public sentiment sentiment analysis block diagram is as shown in Figure 10.
Public sentiment sentiment analysis form is as shown in figure 11.
Public sentiment hot analysis block diagram is as shown in figure 12.
Step 6: result is subjected to visualization processing and preserves and exports with Excel forms
It is realized by key-course ExcelBuild classes and the ExcelUtil classes on backstage and exports data with Excel forms.
Store path is defined, file content is taken out by output character stream, the wiring method of bottom Util is called, will encapsulate
External information well, data message, Table Header information are passed to as three big parameters in Util classes, connect the pass that Build class files are passed to
In three parameters of gauge outfit：File external information is expert at, Table Header information, writes data line by line in output function, finally with
Data are write in the export of Excel forms.
Excel contents after export are as shown in figure 13.
Claims (7)
1. a kind of city management information the analysis of public opinion system, which is characterized in that include the following steps：
Step 1: structure participle database：It is stored simultaneously using based on the database of Oracle to having natural language participle
Database is provided for algorithm calculating to support；
Step 2: realize the acquisition of text data：Using the project management system based on Maven, text data is carried out on foreground
Typing, with Ajax by data storage with Json in url carries out with the interaction on backstage so that server can obtain needs divide
The text message of analysis；
Step 3: emotion pre-processes after Chinese word segmentation and participle：It is carried out substantially by dismembering an ox as skillfully as a butcher algorithm to having text message
Word segmentation processing will segment deposit participle database and put on index and the calculating of emotion value is carried out at the same time during participle, according to
According to set several feature vectors, these feature vectors are divided into positive emotion qualifier and negative emotion qualifier, further according to point
The sentiment analysis of word is calculated and analyzed to be indexed the emotion value of comment or model；
Step 4: it is filtered after participle：It can be referred to as noise vocabulary there are many useless vocabulary in one model, filtration is
By specific algorithm, noise vocabulary is filtered by set keyword feature vector or keyword storehouse, then by the knot after filtering
Fruit is inserted into hot spot dictionary；
Step 5: hot spot and sentiment analysis：The hot spot dictionary that is calculated according to step 4 and each segment corresponding emotion to
Amount carries out the calculating of emotion value and hot statistics to all analysis data, draws the corresponding emotion value of each sentence and the heat of participle
Degree sequence, and analysis result is subjected to visualization processing；
Step 6: result is subjected to visualization processing and preserves and exports with Excel forms.
2. city management information the analysis of public opinion system according to claim 1, which is characterized in that the step 1 is specifically wrapped
Contain：The table design of database based on Oracle, including segmenting table, the carriage after public sentiment table, participle table, keyword table, filtering
The field that feelings table includes has number, content, time, source, sentiment analysis value, the source address of corresponding informance, the participle
The field that table includes has participle number, participle content, the corresponding information source number of participle, part of speech, affection data, source-information
Content, the field that the keyword table includes have keyword number and keyword content, the field that participle table includes after the filtering
Participle numbers, the corresponding information source number of content, participle is segmented after filtering, segments frequency after having filtering.
3. city management information the analysis of public opinion system according to claim 1, which is characterized in that the step two is specific
Including：Model content and number are inputted on foreground by Maven systems, and pass through Json data interchange formats and url and service
Device end interacts, and content of text is passed to server and is preserved according to the data format of public sentiment table.
4. city management information the analysis of public opinion system according to claim 1, which is characterized in that the step three is specific
Comprising：
Using dismembering an ox as skillfully as a butcher, algorithm is segmented：
Character string is passed to algorithm of dismembering an ox as skillfully as a butcher to segment, participle is included with quotation marks after participle, participle is separated with space, will be divided
Word form is rewritten, and is then removed the mark of word segmentation, is left behind participle and space, this result is passed to principal function, principal function calculates
Space number after participle determines participle quantity, removes space, the character string after participle is formally switched in units of participle
Array；
Emotion counts：
Established keyword table and participle table are read, imports positive emotion vector array, negative emotion vector array, emotion degree
Vectorial array and negative emotion vector array, are circulated in for and judge participle vector one by one, emotion vector boolean is become
Amount output, positive emotion true, negative emotion false；
Emotion modification statistics：
The statistical analysis of emotion degree will depending on emotion degree vector, different emotion degree compared with different data,
If centre word be negative word, by binary emotion be set to -1. weak first yuan for negative, then -1 is set to, with second yuan of negative that is multiplied
Negative is front, if there is qualifier before centre word, is multiplied by corresponding weight value to represent emotion degree, first calculates the front of last binary
Degree, then whether determine addition of vectors, draw emotion value result；
Emotion statistical analysis is carried out by triple method, it is 0 to set ternary each several part and total value initial value, note triple difference
For shaping variable res1, res2, res3, note total value is shaping variable res, if centre word is negation words, sets second yuan
Variable res2 is -1, if first yuan is divided into following three kinds of situations for negative：
If 1) first yuan be negative, res1 is set for -1, total value is made to be equal to res2 and is multiplied by res1, the negative of negative is willing
It is fixed；
If 2) there is qualifier before negation words, res1 is assigned a value of corresponding weight value, and res is made to be multiplied by res2 equal to res1, negative to illustrate
Face degree returns to total value res；
If 3) only second yuan of centre word, total value res is equal to the numerical value of centre word res2, returns to total value res；
When centre word res2 is positive emotion, statistical method is with reference to above-mentioned negative statistical analysis；
When centre word is negative word, it is divided into following situation：
1) as only centre word res2, total value res is made to be equal to res2, returns to total value res0；
2) when there are total value res during qualifier res3, is made to be multiplied by res3, return total value res0 equal to res2；
3) when there are during qualifier, make res1 that res be made to be multiplied by res3 equal to res2 and add res1, return to sum equal to corresponding weights
Value res0；
When centre word for conservative negative when, first calculate the positive degree of last binary, then whether determine addition of vectors, draw and most terminate
Fruit；If qualifier will negate vectorial to negate extremely vector and negative multiplication of vectors, its negative degree is drawn.
5. city management information the analysis of public opinion system according to claim 1, which is characterized in that the step four is specific
Comprising：The word that emotion value is 0 is extracted in the participle table drawn by step 3, then the essential information in keyword storehouse
Comparison filters out noise word, and remaining word includes ranking, and including place name, date and name, these words can become hot spot word
It converges, the result after filtering is put into hot spot dictionary, the least unit of statistics is public sentiment member, and public sentiment member is exactly by vocabulary, source
Bind together to be formed a unit insertion hot spot filtering dictionary in, statistics when, if vocabulary source and vocabulary with number
According to having the identical situation of vocabulary in storehouse, then statistics is not counted in, only numerical value adds up in the word frequency for having vocabulary.
6. city management information the analysis of public opinion system according to claim 1, which is characterized in that the step five is specific
Comprising：
Public sentiment sentiment analysis result chart：
There is provided the data comparisons of the sentiment analysis statistics of multi-angle, column map generalization and form in emotion comprehensive statistics module
Corresponding assigned variable is passed to proxy according to foreground selection, obtains backstage Json and shown on foreground, if choosing by output function
Positive Negative sentiments vector can then be shown by trend vector analysis on combobox by selecting, if selection can be shown according to source analysis
Show the site information in various sources；
Public sentiment hot counts：
What public sentiment hot statistics and analysis carried out on the basis of step 3 and step 4, the vocabulary segmented to filtering in table carries out
Extraction, is compared with the information in public sentiment table and carries out further statistical analysis again；
Focus statistics is based on word frequency, using public sentiment source as the unit of statistical analysis member, centre word junction is given to backstage, in proposition
Hot spot vocabulary beyond heart vocabulary forms a unduplicated storage array, then carries out source comparison to the information of extraction, comes
1 is arranged to from the center vocabulary hot spot coefficient in same public feelings information source, the word frequency that hot spot coefficient is multiplied by corresponding centre word i.e. should
The hot spot degree of centre word；If vocabulary to be analyzed is not from same public sentiment source with centre word, the website of word is analysed to
Information is compared with public feelings information storehouse, and the degree of correlation that station external information and centre word are considered to be if inconsistent is zero, such as
Fruit is unanimously then considered as interior information of standing, and closes on algorithm using URL to carry out its calculating with the related coefficient of centre word, finally will
It is exactly its hot spot degree that related coefficient, which is multiplied by its word frequency, and analysis result is exported in units of word.
7. city management information the analysis of public opinion system according to claim 1, which is characterized in that the step six is specific
Comprising：It is realized by key-course ExcelBuild classes and the ExcelUtil classes on backstage and exports data with Excel forms.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711473257.9A CN108052507A (en) | 2017-12-29 | 2017-12-29 | A kind of city management information the analysis of public opinion system and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711473257.9A CN108052507A (en) | 2017-12-29 | 2017-12-29 | A kind of city management information the analysis of public opinion system and method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108052507A true CN108052507A (en) | 2018-05-18 |
Family
ID=62128749
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711473257.9A Pending CN108052507A (en) | 2017-12-29 | 2017-12-29 | A kind of city management information the analysis of public opinion system and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108052507A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109753597A (en) * | 2018-12-31 | 2019-05-14 | 杭州翼兔网络科技有限公司 | A kind of data processing method |
| CN111583363A (en) * | 2020-05-11 | 2020-08-25 | 中国传媒大学 | Visual automatic generation method and system for image-text news |
| CN112418791A (en) * | 2020-11-17 | 2021-02-26 | 北京字跳网络技术有限公司 | Document display method and device and electronic equipment |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102546771A (en) * | 2011-12-27 | 2012-07-04 | 西安博构电子信息科技有限公司 | Cloud mining network public opinion monitoring system based on characteristic model |
| CN104537097A (en) * | 2015-01-09 | 2015-04-22 | 成都布林特信息技术有限公司 | Microblog public opinion monitoring system |
| CN104881417A (en) * | 2014-02-28 | 2015-09-02 | 深圳市网安计算机安全检测技术有限公司 | Public opinion analyzing method and system |
| CN105740228A (en) * | 2016-01-25 | 2016-07-06 | 云南大学 | Internet public opinion analysis method |
| CN106294316A (en) * | 2016-07-29 | 2017-01-04 | 陕西师范大学 | A kind of text emotion based on dictionary analyzes method |
| CN107153658A (en) * | 2016-03-03 | 2017-09-12 | 常州普适信息科技有限公司 | A kind of public sentiment hot word based on weighted keyword algorithm finds method |
-
2017
- 2017-12-29 CN CN201711473257.9A patent/CN108052507A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102546771A (en) * | 2011-12-27 | 2012-07-04 | 西安博构电子信息科技有限公司 | Cloud mining network public opinion monitoring system based on characteristic model |
| CN104881417A (en) * | 2014-02-28 | 2015-09-02 | 深圳市网安计算机安全检测技术有限公司 | Public opinion analyzing method and system |
| CN104537097A (en) * | 2015-01-09 | 2015-04-22 | 成都布林特信息技术有限公司 | Microblog public opinion monitoring system |
| CN105740228A (en) * | 2016-01-25 | 2016-07-06 | 云南大学 | Internet public opinion analysis method |
| CN107153658A (en) * | 2016-03-03 | 2017-09-12 | 常州普适信息科技有限公司 | A kind of public sentiment hot word based on weighted keyword algorithm finds method |
| CN106294316A (en) * | 2016-07-29 | 2017-01-04 | 陕西师范大学 | A kind of text emotion based on dictionary analyzes method |
Non-Patent Citations (1)
| Title |
|---|
| GUANLIN CHEN ET AL: "A Public Opinion Analysis System for Urban Management Information", 《JOURNAL OF COMPUTERS》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109753597A (en) * | 2018-12-31 | 2019-05-14 | 杭州翼兔网络科技有限公司 | A kind of data processing method |
| CN111583363A (en) * | 2020-05-11 | 2020-08-25 | 中国传媒大学 | Visual automatic generation method and system for image-text news |
| CN111583363B (en) * | 2020-05-11 | 2024-05-03 | 中国传媒大学 | Visual automatic generation method and system for graphic news |
| CN112418791A (en) * | 2020-11-17 | 2021-02-26 | 北京字跳网络技术有限公司 | Document display method and device and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103514183B (en) | Information search method and system based on interactive document clustering | |
| CN103544255B (en) | Text semantic relativity based network public opinion information analysis method | |
| CN101957816B (en) | Webpage metadata automatic extraction method and system based on multi-page comparison | |
| CN103324665B (en) | Hot spot information extraction method and device based on micro-blog | |
| CN102930023B (en) | Knowledge based engineering data quality solution | |
| US20050267915A1 (en) | Method and apparatus for recognizing specific type of information files | |
| CN107885499A (en) | A kind of interface document generation method and terminal device | |
| CN104268148B (en) | A kind of forum page Information Automatic Extraction method and system based on time string | |
| CN106815307A (en) | Public Culture knowledge mapping platform and its use method | |
| US10789302B2 (en) | Method and system for extracting user-specific content | |
| US9665561B2 (en) | System and method for performing analysis on information, such as social media | |
| CN107193843A (en) | A kind of character string selection method and device based on AC automatic machines and postfix expression | |
| CN103870506B (en) | Webpage information extraction method and system | |
| CN101593200A (en) | Chinese Web page classification method based on the keyword frequency analysis | |
| CN103324666A (en) | Topic tracing method and device based on micro-blog data | |
| CN110134845A (en) | Project public sentiment monitoring method, device, computer equipment and storage medium | |
| CN107135092A (en) | A kind of Web service clustering method towards global social interaction server net | |
| CN103559234A (en) | System and method for automated semantic annotation of RESTful Web services | |
| CN104331438B (en) | To novel web page contents selectivity abstracting method and device | |
| CN104298732B (en) | The personalized text sequence of network-oriented user a kind of and recommendation method | |
| CN105843796A (en) | Microblog emotional tendency analysis method and device | |
| CN106897309A (en) | The polymerization and device of a kind of similar word | |
| CN105718584B (en) | The method and device that Web page text extracts | |
| CN108052507A (en) | A kind of city management information the analysis of public opinion system and method | |
| Banić et al. | Using big data and sentiment analysis in product evaluation |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20180518 |