CN110162774A - A kind of automation the emotion of news scaling method and device based on financial market market - Google Patents
A kind of automation the emotion of news scaling method and device based on financial market market Download PDFInfo
- Publication number
- CN110162774A CN110162774A CN201910133843.1A CN201910133843A CN110162774A CN 110162774 A CN110162774 A CN 110162774A CN 201910133843 A CN201910133843 A CN 201910133843A CN 110162774 A CN110162774 A CN 110162774A
- Authority
- CN
- China
- Prior art keywords
- news
- market
- history
- correlation
- degree
- 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.)
- Granted
Links
Classifications
-
- 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
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
-
- 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
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/06—Asset management; Financial planning or analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Computational Linguistics (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Development Economics (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Finance (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Accounting & Taxation (AREA)
- Entrepreneurship & Innovation (AREA)
- Databases & Information Systems (AREA)
- Data Mining & Analysis (AREA)
- Game Theory and Decision Science (AREA)
- Human Resources & Organizations (AREA)
- Operations Research (AREA)
- Economics (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- Technology Law (AREA)
- General Business, Economics & Management (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
The present invention relates to a kind of automation the emotion of news scaling method and device based on financial market market, method therein is comprising steps of acquire the history news data of certain period of time；Obtain the corresponding period financial market market historical data of aforementioned certain period of time；Obtain average tendency slope；Each semantic vector of the semantic vector and aforementioned history news that obtain lastest news obtains the correlation score of the forward history news of the M degree of correlation；The M degree of correlation is obtained in the G-bar trend of preceding history news；Demarcate the emotion of news.The beneficial effects of the invention are as follows history news data and historical quotes data is utilized, it is trained to database using LSTM guidance, obtains the degree of correlation and tendency slope of lastest news and history news, and then judge the Sentiment orientation of lastest news；This mode is succinctly feasible, provides advantageous method and apparatus to people.
Description
Technical field
The present invention relates to the crossing domains of e-text and financial industry, more particularly to using history market data, go through
The news data of history demarcates the emotion of lastest news.
Background technique
At home and abroad, media opinion, government policy and mechanism article be quite powerful for financial market or market on
Rise or situation is fallen.People or compassion or happiness are liked summarizing historic market rule, but people are good at amnesia again, unlike electricity
Sub- equipment is the same can effectively trace it is passing, to predict future.Today of picture and text prosperity in internet, financial market market it is pre-
Survey behavior is swept along with news picture and text slowly to grow up together.People know influence of the focus incident for financial market market
Dynamics.
Between the emotion of news and financial market, the research institutions such as university are truly had to make research, but they seem to be biased to learn
Art group goes to develop new method or is integrated to the software of computer storage medium and answer from the angle of algorithm, model, formula mostly
With.The proneness analysis of the emotion of news is necessary, and is no lack of the public opinion air control assistant director of profession in large-scale information mechanism, for monitoring
Instant hot spot or news on network, study and judge the influence to market conditions.However they still lack it is feasible, succinct, high
The technical solution of effect includes method, apparatus to realize the emotion judgement to these lastest news, in following financial market
It makes a profit.
Summary of the invention
The purpose of the present invention is to provide a kind of automation the emotion of news scaling methods and dress based on financial market market
It sets, toolization is in extensive electronic equipment in feasible mode for device here, for sentencing for people to lastest news emotion
It is disconnected.
Automation the emotion of news scaling method provided by the invention based on financial market market, this method includes following step
It is rapid:
(1) the history news data of certain period of time is acquired, the picture and text of policy, finance and economics including mainstream media and picture and text
Issue date；
(2) the corresponding period financial market market historical data of aforementioned certain period of time, including quantizating index are obtained
Numerical value and date；
(3) aforesaid intervals are marked off into N number of section, each history news data in each section is mapped into same zone
Between quantizating index numerical value and the average tendency slope Y on date；
(4) obtain lastest news picture and text include focus incident, the semantic vector after being segmented using lastest news with it is aforementioned
Each semantic vector of history news obtains the correlation score X of the forward history news of the M degree of correlationi, i is 1 to M, and M is less than
Equal to N；
(5) data that above-mentioned each history news data is mapped to the average tendency slope Y on date are obtained according to step (3)
Base relation, G-bar trend Y of the M degree of correlation in preceding history news in obtaining step (4)i, i is 1 to M；
(6) according to the X of acquisitioniAnd YiValue, defines the emotion S=∑ X of the lastest newsiYi, i is 1 to M, if S is greater than 0,
Then the lastest news is front；If S, less than 0, which is negative.
Preferably, mapping relations are obtained in step (3) and use length memory network LSTM mode, and to mapping relations into
Row training.
Preferably, the financial market market can be stock, fund, gold, futures or bond.
It is further preferred that in the step (1) further including the steps that history news data duplicate removal is specially arranged most
Small date intervals threshold value seeks the degree of correlation of content text two-by-two to each history news within minimum date intervals threshold value,
The posterior history news data of issuing time is rejected if the degree of correlation is greater than degree of correlation preset value.
Preferably, the degree of correlation preset value is greater than 70%.
Preferably, the step (2)) in, the financial market market historical data corresponding period, at least compare
The period of the history news data big financial market transactions day.
The present invention also provides a kind of corresponding computer storage medias, are stored in electronic equipment, storage medium operation
The mentioned-above automation news scaling method based on financial market market, and the history news number that there is storage constantly to accumulate
According to the part with financial market market historical data.
Also a kind of automation the emotion of news caliberating device based on financial market market, the device include: the present invention
History news data acquisition unit, for acquiring the history news data of certain period of time, including mainstream media
The issue date of policy, the picture and text of finance and economics and picture and text；
Financial market market historical data acquiring unit, for obtaining the finance of corresponding period of aforementioned certain period of time
Market conditions historical data, including quantizating index numerical value and date；
G-bar trend computing unit goes through each of each section for aforesaid intervals to be marked off N number of section
History news data maps to the quantizating index numerical value in identical section and the average tendency slope Y on date；
Lastest news degree of correlation matching unit, the picture and text for obtaining lastest news include focus incident, using i.e. stylish
Semantic vector after hearing participle obtains the phase of the forward history news of the M degree of correlation with each semantic vector of aforementioned history news
Close degree value Xi, i is 1 to M, and M is less than or equal to N；
The slope tendency acquiring unit of high degree of correlation history news, for being obtained according to aforementioned G-bar trend computing unit
Obtain the database relation that above-mentioned each history news data is mapped to the average tendency slope Y on date, obtaining step lastest news
G-bar trend Y of the M degree of correlation in preceding history news in degree of correlation matching uniti, i is 1 to M；
The emotion of news demarcates unit, for according to the single slope with high degree of correlation history news of lastest news degree of correlation matching
Xi the and Yi value that tendency acquiring unit obtains respectively defines the emotion S=∑ X of the lastest newsiYi, i is 1 to M, if S is greater than 0,
Then the lastest news is front；If S, less than 0, which is negative.
Preferably, the G-bar trend computing unit obtains mapping using length memory network LSTM mode and closes
System, and mapping relations are trained.
Preferably, further include history news data duplicate removal unit, be located in history news data acquisition unit, be used for
Based on minimum date intervals threshold value, content text two-by-two is sought to each history news within minimum date intervals threshold value
The degree of correlation, if the degree of correlation be greater than degree of correlation preset value if reject the posterior history news data of issuing time.
The beneficial effects of the invention are as follows history news data and historical quotes data is utilized, LSTM guidance training is used
At database, the degree of correlation and tendency slope of lastest news and history news are obtained, and then judges the Sentiment orientation of lastest news；
This mode is succinctly feasible, provides advantageous method and apparatus to people.
Detailed description of the invention
Fig. 1 is a kind of automation the emotion of news scaling method step schematic diagram based on financial market market of the invention；
Fig. 2 is that a kind of automation the emotion of news caliberating device based on financial market market of the invention constitutes step signal
One of figure；
Fig. 3 is that a kind of automation the emotion of news caliberating device based on financial market market of the invention constitutes step signal
The two of figure；
Wherein, 0- automates the emotion of news caliberating device, 1- history news data acquisition unit, the financial market 2- market and goes through
History data capture unit, 3- G-bar trend computing unit；4- lastest news degree of correlation matching unit, 5- high degree of correlation history
Slope tendency acquiring unit, 6- the emotion of news calibration unit and the 7- history news data duplicate removal unit of news.
Specific embodiment
In conjunction with attached drawing, the technical schemes of the invention are described in detail further below, but protection scope of the present invention is not limited to
It is described below.
Automation the emotion of news scaling method based on financial market market as shown in Figure 1, this method includes following step
It is rapid:
(1) the history news data of certain period of time is acquired, the picture and text of policy, finance and economics including mainstream media and picture and text
Issue date；
Here it acquires from 2002 most preferably, the Internet media picture and text include emerging in large numbers since media.Period it is more long more
It is good, but consider memory space and processing speed, the period can every year may be used for half a year, 1 year, 3 years or five.In actual storage
It may also include the message comment to picture and text in content, bound together with picture and text, it should be noted that message comment is sometimes right
Sentiment orientation in financial market is bigger；Acquisition mode includes the acquisition of irregular data, can use crawler mode, or
Regular data can be called；
(2) the corresponding period financial market market historical data of aforementioned certain period of time, including quantizating index are obtained
Numerical value and date；
Public opinions such as news picture and text of influence due to to(for) financial market market may have hysteresis quality and extension property, that is, wave
And time range it is larger, therefore the period of the data of financial market market history is not necessarily equal to history news data
Period, but may extend and move back, it will be by embodying in subsequent embodiment.
(3) aforesaid intervals are marked off into N number of section, each history news data in each section is mapped into same zone
Between quantizating index numerical value and the average tendency slope Y on date；
For discontinuous quantizating index numerical value extreme value and other discontinuities, needs to carry out the period isometric or differ
Long division is conducive to mathematically obtain significant trend slope, is that the G-bar trend in section is corresponding to every here
A history news；
(4) obtain lastest news picture and text include focus incident, the semantic vector after being segmented using lastest news with it is aforementioned
Each semantic vector of history news obtains the correlation score X of the forward history news of the M degree of correlationi, i is 1 to M, and M is less than
Equal to N；
In semantic vector process, it is still further preferred that deleting nonsense words therein, enhance this standard for seeking the degree of correlation
True property；
(5) data that above-mentioned each history news data is mapped to the average tendency slope Y on date are obtained according to step (3)
Base relation, G-bar trend Y of the M degree of correlation in preceding history news in obtaining step (4)i, i is 1 to M；
Here the M degree of correlation expresses the content of lastest news in preceding history news, this to replace when data volume is enough
Generation be it is feasible, from the data acquisition test of server, the emotion of this text substitution accuracy rate reaches 78.6%；
(6) according to the X of acquisitioniAnd YiValue, defines the emotion S=∑ X of the lastest newsiYi, i is 1 to M, if S is greater than 0,
Then the lastest news is front；If S, less than 0, which is negative.
Usual degree of correlation X is between 0 to 1, and Y may be positive number, negative or zero, and this succinct product, which is summed, transports
It calculates, is able to ascend efficiency.
In alternative embodiments, mapping relations are obtained in step (3) and use length memory network LSTM mode, and right
Mapping relations are trained.LSTM is the specific type of RNN, repeats neural network and optimizes to mapping relations.
Financial market market can be stock, fund, gold, futures or bond.Other virtual such as bit special purpose currency markets
Field is also feasible.
Have in actual history news the news of short-term issue again or different subjects it is secondary, three times even
Publication, character express therein may be different with sequence, but content is identical.It needs to reject if necessary this to database
Influence.Therefore in the step (1), further include the steps that history news data duplicate removal, the minimum date is specially set
Interval threshold seeks the degree of correlation of content text two-by-two to each history news within minimum date intervals threshold value, if should
The degree of correlation is greater than degree of correlation preset value and then rejects the posterior history news data of issuing time.
The degree of correlation preset value is greater than 70%, is greater than 60% or is greater than 80%, depending on theme phase in picture and text field
With the retransmission probability of news.
It is that the period of the data of financial market market history is not necessarily equal to history news data as previously described
Period, but may extend and move back, therefore in the step (2)) in, the financial market market historical data corresponding time
Section, a financial market transactions day at least bigger than the period of the history news data.
The present invention also provides a kind of corresponding computer storage medias, are stored in electronic equipment, storage medium operation
The automation news scaling method based on financial market market of front, and have the history news data constantly accumulated of storage and
The part of financial market market historical data.Here electronic equipment can for desktop computer, laptop, mobile phone, Ipad or
Other electronic equipments, storing data part can be integrated with computer storage medium, or separately has storage section.
One of them is filled as shown in Fig. 2, the automation the emotion of news based on financial market market is demarcated in embodiment
It sets, which includes:
History news data acquisition unit 1, for acquiring the history news data of certain period of time, including mainstream media
The issue date of policy, the picture and text of finance and economics and picture and text；
Financial market market historical data acquiring unit 2, for obtaining the gold of corresponding period of aforementioned certain period of time
Melt market conditions historical data, including quantizating index numerical value and date；
G-bar trend computing unit 3, for aforesaid intervals to be marked off N number of section, by each of each section
History news data maps to the quantizating index numerical value in identical section and the average tendency slope Y on date；
Lastest news degree of correlation matching unit 4, the picture and text for obtaining lastest news include focus incident, using i.e. stylish
Semantic vector after hearing participle obtains the phase of the forward history news of the M degree of correlation with each semantic vector of aforementioned history news
Close degree value Xi, i is 1 to M, and M is less than or equal to N；
The slope tendency acquiring unit 5 of high degree of correlation history news, for according to aforementioned G-bar trend computing unit
Obtain the database relation that above-mentioned each history news data is mapped to the average tendency slope Y on date, obtaining step, that is, stylish
G-bar trend Y of the M degree of correlation in preceding history news in news degree of correlation matching uniti, i is 1 to M；
The emotion of news demarcates unit 6, for single oblique with high degree of correlation history news according to the matching of the lastest news degree of correlation
Xi the and Yi value that rate tendency acquiring unit obtains respectively defines the emotion S=∑ X of the lastest newsiYi, i is 1 to M, if S is greater than
0, then the lastest news is front；If S, less than 0, which is negative.
With method partial response, G-bar trend computing unit obtains mapping using length memory network LSTM mode and closes
System, and mapping relations are trained.
In another embodiment, as shown in figure 3, further including history news data duplicate removal unit 7, it is located at history news number
According in acquisition unit, it is used for based on minimum date intervals threshold value, it is new to each history within minimum date intervals threshold value
It hears and seeks the degree of correlation of content text two-by-two, reject that issuing time is posterior to be gone through if the degree of correlation is greater than degree of correlation preset value
History news data.
It will be apparent to those skilled in the art that institute's column unit should not be by expression concrete restriction, using the present invention in the present invention
Technical thought or although name is different but completely using listed funtion part in the application, protection of the invention should all be included in
Range.
Claims (10)
1. a kind of automation the emotion of news scaling method based on financial market market, which is characterized in that this method includes following
Step:
(1) the history news data of certain period of time, the publication of the picture and text and picture and text of policy, finance and economics including mainstream media are acquired
Date；
(2) the corresponding period financial market market historical data of aforementioned certain period of time, including quantizating index numerical value are obtained
And the date；
(3) aforesaid intervals are marked off into N number of section, each history news data in each section is mapped into identical section
The average tendency slope Y of quantizating index numerical value and date；
(4) obtaining the picture and text of lastest news includes focus incident, the semantic vector after being segmented using lastest news and aforementioned history
Each semantic vector of news obtains the correlation score X of the forward history news of the M degree of correlationi, i is 1 to M, and M is less than or equal to
N；
(5) it is closed according to the database that step (3) obtain the average tendency slope Y that above-mentioned each history news data is mapped to the date
It is G-bar trend Y of the M degree of correlation in preceding history news in obtaining step (4)i, i is 1 to M；
(6) according to the X of acquisitioniAnd YiValue, defines the emotion S=∑ X of the lastest newsiYi, i is 1 to M, should be i.e. if S is greater than 0
Shi Xinwen is positive face；If S, less than 0, which is negative.
2. the automation the emotion of news scaling method according to claim 1 based on financial market market, which is characterized in that
Mapping relations are obtained in step (3) and use length memory network LSTM mode, and mapping relations are trained.
3. the automation the emotion of news scaling method according to claim 2 based on financial market market, which is characterized in that
The financial market market can be stock, fund, gold, futures or bond.
4. the automation the emotion of news scaling method according to claim 1 based on financial market market, it is characterised in that:
In the step (1), further include the steps that history news data duplicate removal, minimum date intervals threshold value is specially set, to most
Each history news within small date intervals threshold value seeks the degree of correlation of content text two-by-two, if the degree of correlation is greater than correlation
Degree preset value then rejects the posterior history news data of issuing time.
5. the automation the emotion of news scaling method according to claim 4 based on financial market market, which is characterized in that
The degree of correlation preset value is greater than 70%.
6. the automation news scaling method according to claim 1 based on financial market market, it is characterised in that: described
Step (2)) in, the financial market market historical data corresponding period, at least than the history news data when
Between Duan great Yi financial market transactions day.
7. a kind of computer storage media, is stored in electronic equipment, which is characterized in that the storage medium runs claim 1
To the automation news scaling method described in one of 6 based on financial market market, and the history that there is storage constantly to accumulate is new
Hear the part of data and financial market market historical data.
8. a kind of automation the emotion of news caliberating device based on financial market market, which is characterized in that the device includes:
History news data acquisition unit (1), for acquiring the history news data of certain period of time, the political affairs including mainstream media
The issue date of plan, the picture and text of finance and economics and picture and text；
Financial market market historical data acquiring unit (2), for obtaining the finance of corresponding period of aforementioned certain period of time
Market conditions historical data, including quantizating index numerical value and date；
G-bar trend computing unit (3) goes through each of each section for aforesaid intervals to be marked off N number of section
History news data maps to the quantizating index numerical value in identical section and the average tendency slope Y on date；
Lastest news degree of correlation matching unit (4), the picture and text for obtaining lastest news include focus incident, using lastest news
Semantic vector after participle obtains the related of the forward history news of the M degree of correlation to each semantic vector of aforementioned history news
Degree value Xi, i is 1 to M, and M is less than or equal to N；
The slope tendency acquiring unit (5) of high degree of correlation history news, for being obtained according to aforementioned G-bar trend computing unit
Obtain the database relation that above-mentioned each history news data is mapped to the average tendency slope Y on date, obtaining step lastest news
G-bar trend Y of the M degree of correlation in preceding history news in degree of correlation matching uniti, i is 1 to M；
The emotion of news demarcates unit (6), for according to lastest news degree of correlation matching unit (4) and high degree of correlation history news
Xi the and Yi value that slope tendency acquiring unit (5) obtains respectively defines the emotion S=∑ X of the lastest newsiYi, i is 1 to M, if
S is greater than 0, then the lastest news is front；If S, less than 0, which is negative.
9. the automation the emotion of news caliberating device according to claim 8 based on financial market market, which is characterized in that
The G-bar trend computing unit (3) obtains mapping relations using length memory network LSTM mode, and to mapping relations
It is trained.
10. the automation the emotion of news caliberating device according to claim 9 based on financial market market, feature exist
In further including history news data duplicate removal unit (7), be located in history news data acquisition unit (1), for minimum day
Based on phase interval threshold, the correlation of content text two-by-two is sought to each history news within minimum date intervals threshold value
Degree rejects the posterior history news data of issuing time if the degree of correlation is greater than degree of correlation preset value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910133843.1A CN110162774B (en) | 2019-02-22 | 2019-02-22 | Automatic news emotion calibration method and device based on financial market quotation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910133843.1A CN110162774B (en) | 2019-02-22 | 2019-02-22 | Automatic news emotion calibration method and device based on financial market quotation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110162774A true CN110162774A (en) | 2019-08-23 |
| CN110162774B CN110162774B (en) | 2022-12-23 |
Family
ID=67645390
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910133843.1A Active CN110162774B (en) | 2019-02-22 | 2019-02-22 | Automatic news emotion calibration method and device based on financial market quotation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110162774B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111310709A (en) * | 2020-03-02 | 2020-06-19 | 邓谊 | Image-text annual newspaper emotion calibration method and system |
| CN114036208A (en) * | 2021-11-09 | 2022-02-11 | 建信金融科技有限责任公司 | Model training and sensitivity analysis method, device, equipment and medium |
| CN114386433A (en) * | 2022-01-12 | 2022-04-22 | 中国农业银行股份有限公司 | Data processing method, device and equipment based on emotion analysis and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030187772A1 (en) * | 2002-01-18 | 2003-10-02 | Ron Papka | System and method for predicting security price movements using financial news |
| CN103778215A (en) * | 2014-01-17 | 2014-05-07 | 北京理工大学 | Stock market forecasting method based on sentiment analysis and hidden Markov fusion model |
| CN103985055A (en) * | 2014-05-30 | 2014-08-13 | 西安交通大学 | Stock market investment decision-making method based on network analysis and multi-model fusion |
| CN104751363A (en) * | 2015-03-24 | 2015-07-01 | 北京工商大学 | Stock medium and long term trend prediction method and system based on Bayes classifier |
| US20170018033A1 (en) * | 2015-07-15 | 2017-01-19 | Foundation Of Soongsil University Industry Cooperation | Stock fluctuatiion prediction method and server |
| AU2018101524A4 (en) * | 2018-10-14 | 2018-11-15 | Chai, Xiayun MISS | Stock prediction research based on finiancial news by svm |
-
2019
- 2019-02-22 CN CN201910133843.1A patent/CN110162774B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030187772A1 (en) * | 2002-01-18 | 2003-10-02 | Ron Papka | System and method for predicting security price movements using financial news |
| CN103778215A (en) * | 2014-01-17 | 2014-05-07 | 北京理工大学 | Stock market forecasting method based on sentiment analysis and hidden Markov fusion model |
| CN103985055A (en) * | 2014-05-30 | 2014-08-13 | 西安交通大学 | Stock market investment decision-making method based on network analysis and multi-model fusion |
| CN104751363A (en) * | 2015-03-24 | 2015-07-01 | 北京工商大学 | Stock medium and long term trend prediction method and system based on Bayes classifier |
| US20170018033A1 (en) * | 2015-07-15 | 2017-01-19 | Foundation Of Soongsil University Industry Cooperation | Stock fluctuatiion prediction method and server |
| AU2018101524A4 (en) * | 2018-10-14 | 2018-11-15 | Chai, Xiayun MISS | Stock prediction research based on finiancial news by svm |
Non-Patent Citations (1)
| Title |
|---|
| 陈海文等: "应用财经新闻挖掘的金融品种价格走势预测", 《计算机工程与科学》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111310709A (en) * | 2020-03-02 | 2020-06-19 | 邓谊 | Image-text annual newspaper emotion calibration method and system |
| CN114036208A (en) * | 2021-11-09 | 2022-02-11 | 建信金融科技有限责任公司 | Model training and sensitivity analysis method, device, equipment and medium |
| CN114386433A (en) * | 2022-01-12 | 2022-04-22 | 中国农业银行股份有限公司 | Data processing method, device and equipment based on emotion analysis and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110162774B (en) | 2022-12-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20180182038A1 (en) | System for generating data from social media messages for the real-time evaluation of publicly traded assets | |
| Sun et al. | Predicting stock price returns using microblog sentiment for chinese stock market | |
| Chen | Business and market intelligence 2.0, Part 2 | |
| CN110162774A (en) | A kind of automation the emotion of news scaling method and device based on financial market market | |
| Garcia-Lopez et al. | Analysis of relationships between tweets and stock market trends | |
| Rouhani et al. | Crypto-currencies narrated on tweets: a sentiment analysis approach | |
| Svetlov et al. | Sentiment analysis of posts and comments in the accounts of Russian politicians on the social network | |
| CN112419029B (en) | Similar financial institution risk monitoring method, risk simulation system and storage medium | |
| Chatterjee et al. | Investor classification and sentiment analysis | |
| US20140013223A1 (en) | System and method for contextual visualization of content | |
| Banwo | Artificial intelligence and financial services: Regulatory tracking and change management | |
| Sigari et al. | Analyzing the past, improving the future: a multiscale opinion tracking model for optimizing business performance | |
| Chao et al. | Opinion mining and the visualization of stock selection in quantitative trading | |
| Kennis | Multi-channel discourse as an indicator for Bitcoin price and volume movements | |
| Chen | Designing an Investment Research System for Asset Management Based on Natural Language Processing | |
| Sasaki et al. | Evaluation of VI Index Forecasting Model by Machine Learning for Yahoo! Stock BBS Using Volatility Trading Simulation. | |
| Devi et al. | Social media aided sentiment analysis for stock prediction | |
| Manzoor et al. | Stock exchange prediction using financial news and sentiment analysis | |
| Boutsoukis | Near Real-Time Cryptocurrency Sentiment Analysis | |
| MONESA et al. | SENTIMENT ANALYSIS ON INVESTMENT TOPIC IN INDONESIA USING MACHINE LEARNING ALGORITHMS APPROACH | |
| Rateb et al. | Cryptocurrency Price Prediction: A Comparative Sentiment Analysis Approach Using SVM, CNN-LSTM, and Pysentimento during Times of Crisis | |
| Bhatia et al. | Machine Learning Approach to Extracting Emotions Information from Open Source Data for Relative Forecasting of Stock Prices | |
| Rateb et al. | Predicting Cryptocurrency Prices during Periods of Conflict: A Comparative Sentiment Analysis Approach Using SVM, CNN-LSTM, and PySentimento | |
| Rateb et al. | Cryptocurrency Price Forecasting Implementing Sentiment Analysis during the Russian-Ukrainian War | |
| Rateb et al. | Sentiment Analysis in Times of Crisis: Evaluating SVM, CNN-LSTM, and Psyntimiento |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |