CN109190105A - A kind of enterprise's public sentiment macroscopic view sentiment analysis method - Google Patents
A kind of enterprise's public sentiment macroscopic view sentiment analysis method Download PDFInfo
- Publication number
- CN109190105A CN109190105A CN201810688896.5A CN201810688896A CN109190105A CN 109190105 A CN109190105 A CN 109190105A CN 201810688896 A CN201810688896 A CN 201810688896A CN 109190105 A CN109190105 A CN 109190105A
- Authority
- CN
- China
- Prior art keywords
- enterprise
- value
- emotion
- threshold value
- article
- 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
- 238000004458 analytical method Methods 0.000 title claims abstract description 19
- 230000008451 emotion Effects 0.000 claims abstract description 71
- 238000000034 method Methods 0.000 claims abstract description 13
- 238000004364 calculation method Methods 0.000 claims abstract description 5
- 230000006870 function Effects 0.000 claims description 5
- 238000010801 machine learning Methods 0.000 claims description 2
- 230000007935 neutral effect Effects 0.000 claims description 2
- 230000000694 effects Effects 0.000 claims 1
- 241001062009 Indigofera Species 0.000 description 7
- 230000007423 decrease Effects 0.000 description 2
- 230000002996 emotional effect Effects 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/205—Parsing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
Abstract
The present invention discloses a kind of enterprise's public sentiment macroscopic view sentiment analysis method.The method comprehensively considers document content, the correlation of document and enterprise, deliver the authority of the media of document, document delivers the key factors such as time, different calculating weights is assigned to each document respectively, pass through COMPREHENSIVE CALCULATING formula, calculate overall macroscopical emotion score value of enterprise's public sentiment, and macroscopical emotion score value is mapped to [- 1, 1] section, set two score value thresholdings, that is the first threshold value and the second threshold value, emotion score value is qualitative for passiveness lower than the first threshold value, emotion score value is qualitative for neutrality between the first threshold value and the second threshold value, it is positive that emotion score value, which is higher than the qualitative of the second threshold value,.Public sentiment emotion tendency on the day of the comprehensive a variety of element factor calculation enterprises of the method for the invention, and not will use the current share price of enterprise and analyzed, the relationship that can be more accurately fitted between emotion and share price.
Description
Technical field
The invention belongs to natural language processing technique fields, and in particular to a kind of enterprise's public sentiment macroscopic view sentiment analysis method.
Background technique
It is more and more about all kinds of news of enterprise, evaluation article on network with popularizing for network application, while each
In class social media, online friend is also more and more abundant about all kinds of comment informations of enterprise.In order to analyze society comprehensively to enterprise
Attention rate, the degree of recognition of industry, need the text information being present in network to these to be analyzed and processed, are wrapped according in document
The related emotion information contained, comprehensive analysis masses predict the developing prospect of enterprise to the attitude and emotion of enterprise.
The currently used method for asking whole emotion distribution is to take the mode averaged based on statistical result.That is needle
To each listed company, every text can be given by having many relevant news informations and social data by sentiment analysis algorithm
Chapter calculates emotional value, and emotional value includes three classification such as positive, neutrality, passiveness.When the whole feelings that calculate a company day
When feeling tendency, general way is to seek the average value of the same day all articles.
It is this to judge public to ask the method for the overall emotion of enterprise there are following by way of pleading and feeling average value
Topic: first is that the factor considered is not complete, such as the correlation of document, the authorities of the media for delivering document, document are not accounted for
The key factors such as time are delivered, emotion score value is only obtained according to the content of document, limitation is obvious；Second is that in the presence of just forsaking one's love
The case where sense is offset cannot really reflect public emotion panorama, cause sentiment analysis inaccurate.
Summary of the invention
Incomplete in order to solve the consideration of element existing for existing analysis enterprise public sentiment macroscopic view emotion algorithm, there are positive negative affects
It offsets, it is difficult to which the problem of reflecting people's emotions overall picture, the present invention provide a kind of enterprise's public sentiment macroscopic view sentiment analysis method, the side
Method comprehensively considers document content, the correlation of document and enterprise, the authority of media for delivering document, document and delivers time etc.
Key factor assigns different calculating weights to each document respectively, and by COMPREHENSIVE CALCULATING formula, the totality for calculating enterprise's public sentiment is macro
See emotion score value.
To realize above-mentioned target, the invention adopts the following technical scheme:
A kind of enterprise's public sentiment macroscopic view sentiment analysis method, the method comprehensively consider the phase of document content, document and enterprise
Guan Xing, the authority of media for delivering document, document deliver the key factors such as time, respectively assign different meters to each document
Weight is calculated, by COMPREHENSIVE CALCULATING formula, calculates overall macroscopical emotion score value of enterprise's public sentiment, and macroscopical emotion score value is mapped to
[- 1,1] section, sets two score value thresholdings, i.e. the first threshold value and the second threshold value, and emotion score value is lower than the first threshold value
It is qualitative for passiveness, by emotion score value between the first threshold value and the second threshold value it is qualitative for neutrality, by emotion score value
Qualitative higher than the second threshold value is positive.
Shown in the COMPREHENSIVE CALCULATING formula such as formula (1).
In formula:
Si: indicate the emotion score value of single document, the emotion of article expression is more positive, and the value is bigger；Otherwise it is smaller.
Mi: the weight of media where indicating single article, the media published an article are more important, and the value is bigger；Otherwise it is smaller.
Ri: the correlation coefficient of single article is indicated, article is more related to event, and the value is bigger；
Otherwise it is smaller.
e-λ(t-ti): the time dough softening, article deliver that the time is closer from now, and the value is bigger；Otherwise it is smaller.T indicates the same day,
tiIndicate the date that article is delivered.
N indicates the total amount of all kinds of articles relevant to enterprise in 24 hours in the past.
The default value of the adjustment factor α be 1, controlled by constantly adjusting the adjustment factor fitting data area and
Smoothness.
Final calculating score is less than -0.2 and is classified as Negative Affect, and between -0.2 to 0.2 is neutral emotion, is greater than
0.2 is positive emotion.
The emotion score value S of the single documentiCalculation method be using machine learning algorithm, using maximum entropy model into
Row study obtains.
The calculation method of the weight Mi of media is to be given a mark to obtain according to the influence power of media where the single article
's.
The correlation coefficient R of the single articleiIt is to be realized using TD-IDF algorithm.
A kind of enterprise's public sentiment macroscopic view sentiment analysis method, the described method comprises the following steps:
1) all kinds of articles relevant to enterprise in 24 hours are collected；
2) the emotion score value of every article is calculated separately；
3) weight for delivering the media of every article is calculated separately；
4) related coefficient of every article and enterprise is calculated separately；
5) adjustment factor α is determined；
6) tanh () function is determined；
7) emotion score value thresholding 1 and thresholding 2 are determined；
7) the public sentiment macroscopic view emotion score value of enterprise is calculated using formula (1)；
8) qualitative to the public sentiment macroscopic view emotion of enterprise according to the first threshold value of emotion score value and the second threshold value.
The advantages and benefits of the present invention are: having comprehensively considered the feelings of the relevant news of daily enterprise, social article
Sense trend, in conjunction with the issuing time of article, the factors such as the weight of the publication medium of article, the emotion on the day of COMPREHENSIVE CALCULATING is walked
Gesture, and not will use current share price and analyzed, the relationship that can be more accurately fitted between emotion and share price.
Detailed description of the invention
Attached drawing 1 is the fitted figure that the embodiment of the present invention 1 is directed to McDonald Co., Ltd daily share price and emotion tendency.It is bent in figure
Line is the fluctuation of public sentiment macroscopic view emotion, and dash area is volatility.
Attached drawing 2 is the embodiment of the present invention 2 for the emotion of reddish yellow indigo plant kindergarten and the fitted figure of share price.In figure, curve is
The fluctuation of public sentiment macroscopic view emotion, dash area is volatility.
Specific embodiment
Below with reference to embodiment, the invention will be further described.
Embodiment 1
Referring to attached drawing 1.The public sentiment macroscopic view emotion score value of McDonald Co., Ltd is calculated, and emotion fluctuation is carried out with volatility
Fitting.It executes according to the following steps.
1) all kinds of articles relevant to McDonald Co., Ltd in 24 hours are collected；
2) the emotion score value of every article is calculated separately；
3) weight for delivering the media of every article is calculated separately；
4) related coefficient of every article and McDonald Co., Ltd is calculated separately；
5) adjustment factor α is determined；
6) tanh () function is determined；
7) the first threshold value of emotion score value and the second threshold value are determined, respectively -0.2 and 0.2；
7) the public sentiment macroscopic view emotion score value of McDonald Co., Ltd is calculated using formula (1)；
8) qualitative to the public sentiment macroscopic view emotion of McDonald Co., Ltd according to the first threshold value of emotion score value and the second threshold value；
9) the public sentiment macroscopic view emotion of McDonald Co., Ltd is fluctuated and is fitted with volatility.
By attached drawing 1 it is found that the method through the invention, can react the degree of fitting of share price and emotion on the whole, work as feelings
When sense score is high, share price also will rise accordingly, and when emotion score is low, share price can also decline accordingly.
Embodiment 2
Referring to attached drawing 2.Calculate reddish yellow indigo plant kindergarten public sentiment macroscopic view emotion score value, and by emotion fluctuation and volatility into
Row fitting.It executes according to the following steps.
1) all kinds of articles relevant to reddish yellow indigo plant kindergarten in 24 hours are collected；
2) the emotion score value of every article is calculated separately；
3) weight for delivering the media of every article is calculated separately；
4) related coefficient of every article Yu reddish yellow indigo plant kindergarten is calculated separately；
5) adjustment factor α is determined；
6) tanh () function is determined；
7) the first threshold value of emotion score value and the second threshold value are determined, respectively -0.2 and 0.2；
7) the public sentiment macroscopic view emotion score value of reddish yellow indigo plant kindergarten is calculated using formula (1)；
8) fixed to the public sentiment macroscopic view emotion of reddish yellow indigo plant kindergarten according to the first threshold value of emotion score value and the second threshold value
Property；
9) the public sentiment macroscopic view emotion of reddish yellow indigo plant kindergarten is fluctuated and is fitted with volatility.
By attached drawing 2 it is found that the method through the invention, can react the degree of fitting of share price and emotion on the whole, work as feelings
When sense score is high, share price also will rise accordingly, and when emotion score is low, share price can also decline accordingly.Box
The emotion fluctuation of institute's covering part is very fierce, and correspondingly, volatility is also very violent.
Finally, it should be noted that obviously, the above embodiment is merely an example for clearly illustrating the present invention, and simultaneously
The non-restriction to embodiment.For those of ordinary skill in the art, it can also do on the basis of the above description
Other various forms of variations or variation out.There is no necessity and possibility to exhaust all the enbodiments.And thus drawn
The obvious changes or variations of stretching are still in the protection scope of this invention.
Claims (8)
1. a kind of enterprise's public sentiment macroscopic view sentiment analysis method, it is characterised in that: the method comprehensively consider document content, document with
The correlation of enterprise, the authority of media for delivering document, document deliver time factor, assign respectively to each document different
Weight is calculated, by COMPREHENSIVE CALCULATING formula, calculates overall macroscopical emotion score value of enterprise's public sentiment, and macroscopical emotion score value is mapped
To [- 1,1] section, two score value thresholdings, i.e. the first threshold value and the second threshold value are set, emotion score value is lower than the first thresholding
The qualitative of value is passiveness, and emotion score value is qualitative to be neutral between the first threshold value and the second threshold value, by emotion point
It is positive that value, which is higher than the qualitative of the second threshold value,.
2. a kind of enterprise's public sentiment macroscopic view sentiment analysis method according to claim 1, which is characterized in that the COMPREHENSIVE CALCULATING
Formula is as follows:
In formula:
Si: indicate the emotion score value of single document, the emotion of article expression is more positive, and the value is bigger；Otherwise it is smaller；
Mi: the weight of media where indicating single article, the media published an article are more important, and the value is bigger；Otherwise it is smaller；
Ri: indicate the correlation coefficient of single article, article is more related to event, and the value is bigger；Otherwise it is smaller；
The time dough softening, article deliver that the time is closer from now, and the value is bigger；Otherwise smaller, t indicates the same day, tiTable
Show the date that article is delivered；
N: the total amount of all kinds of articles relevant to enterprise in 24 hours in the past is indicated；
Tanh () function: effect is will to refer to that target value is mapped to [- 1,1] section；
α: adjustment factor.
3. a kind of enterprise's public sentiment macroscopic view sentiment analysis method according to claim 1, it is characterised in that: first thresholding
Value is -0.2, and second threshold value is 0.2.
4. a kind of enterprise's public sentiment macroscopic view sentiment analysis method according to claim 1 or 2, it is characterised in that: described single
The emotion score value S of documentiCalculation method be to be learnt using maximum entropy model using machine learning algorithm.
5. a kind of enterprise's public sentiment macroscopic view sentiment analysis method according to claim 1 or 2, it is characterised in that: described single
The weight M of media where articleiCalculation method be to be given a mark according to the influence power of media.
6. a kind of enterprise's public sentiment macroscopic view sentiment analysis method according to claim 1 or 2, it is characterised in that: described single
The correlation coefficient R of articleiIt is to be realized using TD-IDF algorithm.
7. a kind of enterprise's public sentiment macroscopic view sentiment analysis method according to claim 1 or 2, it is characterised in that: the adjusting
The default value of factor alpha is 1, and the data area and smoothness of fitting are controlled by constantly adjusting the adjustment factor.
8. a kind of enterprise's public sentiment macroscopic view sentiment analysis method according to claim 1, which is characterized in that the method includes
Following steps:
1) all kinds of articles relevant to enterprise in 24 hours are collected；
2) the emotion score value of every article is calculated separately；
3) weight for delivering the media of every article is calculated separately；
4) related coefficient of every article and enterprise is calculated separately；
5) adjustment factor α is determined；
6) tanh () function is determined；
7) the first threshold value of emotion score value and the second threshold value are determined；
7) the public sentiment macroscopic view emotion score value of enterprise is calculated using COMPREHENSIVE CALCULATING formula；
8) qualitative to the public sentiment macroscopic view emotion of enterprise according to the first threshold value of emotion score value and the second threshold value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810688896.5A CN109190105A (en) | 2018-06-28 | 2018-06-28 | A kind of enterprise's public sentiment macroscopic view sentiment analysis method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810688896.5A CN109190105A (en) | 2018-06-28 | 2018-06-28 | A kind of enterprise's public sentiment macroscopic view sentiment analysis method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109190105A true CN109190105A (en) | 2019-01-11 |
Family
ID=64948617
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810688896.5A Pending CN109190105A (en) | 2018-06-28 | 2018-06-28 | A kind of enterprise's public sentiment macroscopic view sentiment analysis method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109190105A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111414455A (en) * | 2020-03-20 | 2020-07-14 | 北京百度网讯科技有限公司 | Public opinion analysis method, device, electronic equipment and readable storage medium |
| CN111950860A (en) * | 2020-07-21 | 2020-11-17 | 中证征信（深圳）有限公司 | Method and device for monitoring enterprise public opinion risk index |
| CN112686048A (en) * | 2020-12-23 | 2021-04-20 | 沈阳新松机器人自动化股份有限公司 | Emotion recognition method and device based on fusion of voice, semantics and facial expressions |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104462487A (en) * | 2014-12-19 | 2015-03-25 | 南开大学 | Individualized online news comment mood forecast method capable of fusing multiple information sources |
| CN106294532A (en) * | 2016-05-18 | 2017-01-04 | 广东电网有限责任公司信息中心 | The image appraisal algorithm analyzed based on microblog emotional |
| CN107045497A (en) * | 2017-05-04 | 2017-08-15 | 成都华栖云科技有限公司 | A kind of quick newsletter archive content sentiment analysis system and method |
| CN107633044A (en) * | 2017-09-14 | 2018-01-26 | 国家计算机网络与信息安全管理中心 | A kind of public sentiment knowledge mapping construction method based on focus incident |
| CN107705005A (en) * | 2017-09-27 | 2018-02-16 | 吴殿义 | A kind of movie and television contents Valuation Method |
-
2018
- 2018-06-28 CN CN201810688896.5A patent/CN109190105A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104462487A (en) * | 2014-12-19 | 2015-03-25 | 南开大学 | Individualized online news comment mood forecast method capable of fusing multiple information sources |
| CN106294532A (en) * | 2016-05-18 | 2017-01-04 | 广东电网有限责任公司信息中心 | The image appraisal algorithm analyzed based on microblog emotional |
| CN107045497A (en) * | 2017-05-04 | 2017-08-15 | 成都华栖云科技有限公司 | A kind of quick newsletter archive content sentiment analysis system and method |
| CN107633044A (en) * | 2017-09-14 | 2018-01-26 | 国家计算机网络与信息安全管理中心 | A kind of public sentiment knowledge mapping construction method based on focus incident |
| CN107705005A (en) * | 2017-09-27 | 2018-02-16 | 吴殿义 | A kind of movie and television contents Valuation Method |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111414455A (en) * | 2020-03-20 | 2020-07-14 | 北京百度网讯科技有限公司 | Public opinion analysis method, device, electronic equipment and readable storage medium |
| CN111414455B (en) * | 2020-03-20 | 2024-03-01 | 北京百度网讯科技有限公司 | Public opinion analysis method, public opinion analysis device, electronic equipment and readable storage medium |
| CN111950860A (en) * | 2020-07-21 | 2020-11-17 | 中证征信（深圳）有限公司 | Method and device for monitoring enterprise public opinion risk index |
| CN111950860B (en) * | 2020-07-21 | 2024-04-16 | 中证征信（深圳）有限公司 | Monitoring method and device for enterprise public opinion risk index |
| CN112686048A (en) * | 2020-12-23 | 2021-04-20 | 沈阳新松机器人自动化股份有限公司 | Emotion recognition method and device based on fusion of voice, semantics and facial expressions |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Bursztyn et al. | Opinions as facts | |
| Kletzer | Job loss from imports: Measuring the costs | |
| Hahn et al. | Efficiency and equity impacts of energy subsidies | |
| CN109190105A (en) | A kind of enterprise's public sentiment macroscopic view sentiment analysis method | |
| CN109190121A (en) | Car review sentiment analysis method based on automobile body and part-of-speech rule | |
| He et al. | The impact of local government investment on the carbon emissions reduction effect: An empirical analysis of panel data from 30 provinces and municipalities in China | |
| Occhino | The flattening of the Phillips curve: Policy implications depend on the cause | |
| CN107045533A (en) | Educational resource based on label recommends method and system | |
| CN109902229A (en) | A kind of interpretable recommended method based on comment | |
| Ahn et al. | Understanding the performance of collaborative filtering recommendation through social network analysis | |
| Chen et al. | A dynamic analysis of the relationship between investor sentiment and stock market realized volatility: Evidence from China | |
| CN106846029A (en) | Collaborative Filtering Recommendation Algorithm based on genetic algorithm and new similarity calculative strategy | |
| CN104866490B (en) | A kind of video intelligent recommended method and its system | |
| Cipollone et al. | Women's employment: joining explanations based on individual characteristics and on contextual factors | |
| Satish et al. | A case study of barriers and drivers for energy efficiency in an Indian city | |
| Wang et al. | Monetary policy rules under a fixed exchange rate regime: empirical evidence from China | |
| Clegg | Taking one for the team: Partisan alignment and planning outcomes in England | |
| De Gispert | The economic analysis of industrial water demand: A review | |
| Timmaraju et al. | Towards Fairness in Personalized Ads Using Impression Variance Aware Reinforcement Learning | |
| Zhang et al. | Collaborative filtering algorithm based on improved time function and user similarity | |
| Born et al. | Monetary Policy in the Age of Social Media: A Twitter-Based Inflation Analysis | |
| Wang et al. | Building consumer confidence index based on social media big data | |
| Zhang et al. | Application of clustering algorithm on TV programmes preference grouping of subscribers | |
| Fan et al. | Local education expenditures and educational inequality in China | |
| Medyawati et al. | Factors Influencing the Demand for Indonesian Tourism |
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: 20190111 |