CN110728140A - Emotion analysis and theme feature-based emergent event public opinion evolution analysis method - Google Patents
Emotion analysis and theme feature-based emergent event public opinion evolution analysis method Download PDFInfo
- Publication number
- CN110728140A CN110728140A CN201810718522.3A CN201810718522A CN110728140A CN 110728140 A CN110728140 A CN 110728140A CN 201810718522 A CN201810718522 A CN 201810718522A CN 110728140 A CN110728140 A CN 110728140A
- Authority
- CN
- China
- Prior art keywords
- emergency
- topic
- evolution
- emotion
- text
- 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
- 230000008451 emotion Effects 0.000 title claims abstract description 46
- 238000004458 analytical method Methods 0.000 title claims abstract description 37
- 238000000034 method Methods 0.000 claims abstract description 62
- 230000011218 segmentation Effects 0.000 claims abstract description 25
- 230000008569 process Effects 0.000 claims abstract description 23
- 230000008859 change Effects 0.000 claims abstract description 20
- 238000011161 development Methods 0.000 claims abstract description 16
- 230000018109 developmental process Effects 0.000 claims abstract description 16
- 238000004364 calculation method Methods 0.000 claims abstract description 11
- 230000002996 emotional effect Effects 0.000 claims description 12
- 238000012545 processing Methods 0.000 claims description 6
- 230000001788 irregular Effects 0.000 claims description 2
- 238000012163 sequencing technique Methods 0.000 claims description 2
- 238000009826 distribution Methods 0.000 description 42
- 238000005070 sampling Methods 0.000 description 19
- 241000907316 Zika virus Species 0.000 description 9
- 238000005516 engineering process Methods 0.000 description 6
- 230000014509 gene expression Effects 0.000 description 3
- 238000003058 natural language processing Methods 0.000 description 3
- 230000033228 biological regulation Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000005065 mining Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000007480 spreading Effects 0.000 description 2
- 238000007476 Maximum Likelihood Methods 0.000 description 1
- 241000700605 Viruses Species 0.000 description 1
- 230000002776 aggregation Effects 0.000 description 1
- 238000004220 aggregation Methods 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000013467 fragmentation Methods 0.000 description 1
- 238000006062 fragmentation reaction Methods 0.000 description 1
- 230000000977 initiatory effect Effects 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 230000010287 polarization Effects 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000002195 synergetic effect Effects 0.000 description 1
- 238000012731 temporal analysis Methods 0.000 description 1
- 238000000700 time series analysis Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
Images
Classifications
-
- 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
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/01—Social networking
-
- 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
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Tourism & Hospitality (AREA)
- Primary Health Care (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Health & Medical Sciences (AREA)
- Strategic Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Economics (AREA)
- Computing Systems (AREA)
- Development Economics (AREA)
- Educational Administration (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses an analysis method for public sentiment evolution of emergencies based on sentiment analysis and topic characteristics, which mainly comprises the following steps: firstly, according to the characteristics of Chinese word segmentation, a word segmentation method based on character string matching is adopted to segment a data text; secondly, expanding the weight and the characteristics of the segmented words by using synonym forest corpus and carrying out emotion value calculation; then calculating and time-slicing the data based on the emotion difference value; and finally, the data after time slicing is used as the input of a Dynamic Topic Model (DTM), so that the change and the heat of the subject word of the emergency on a time sequence are obtained, the development trend of the emergency is predicted, and the role of emotion analysis in the topic evolution process is verified by contrasting the topic evolution result only considering text similarity slicing.
Description
Technical Field
The invention belongs to the field of natural language processing, and particularly relates to an emotion analysis and theme feature emergent event public opinion demonstration analysis method, which is used for mining emotions expressed by netizens under an emergent event.
Background
At present, in the conventional method for researching the change of observation values of the number of comments on the public sentiment aspect to obtain the public sentiment development situation, although the time series analysis directly takes the observation data corresponding to the states of the object at different moments as research objects, a time series model is established by analyzing and researching the characteristics of the time series data, and then the change rule of the object along with the time development is discovered and described in some way on the basis of the fitted model. For the analysis of time series, researchers have accumulated some classical methods, mainly including model methods, index methods and graph methods. However, the method only reflects the overall development trend of the emergency, and deep and dynamic rules are difficult to find.
In addition, different themes often appear in the whole process of spreading the emergency, emotion expressions of netizens under different themes also directly influence the spreading speed and trend of the emergency, and digging aiming at the theme of the microblog and the corresponding emotion of each theme can help organizations such as governments, enterprises and the like to quickly know public emotion trend and proportion change in the process of occurrence of the emergency, so that the emergency is predicted and regulated.
Therefore, on the basis, a network public opinion evolution analysis method based on topic analysis is evolved, and a topic discovery method mostly adopts a method for inspecting the scale of related comments on the basis of clustering, so that the emotional factors of netizen comments are often not sufficiently concerned, and the topic discovery method usually stays at the discovery degree of hot topics, and is not further distinguished according to the tendency characteristics. In recent years, although scholars have conducted deep analysis research on public sentiment characteristics and given some definitions of public sentiment key points, specific mathematical models are not given, and feasible discovery schemes are not provided, and most of the discovery methods are manual collection and arrangement. The emotion of netizens can be analyzed in a fine granularity mode, the proportion change and the trend of the emotion of the public sentiment in the occurrence process can be mastered, and the method has important practical significance for prediction and regulation of the public sentiment.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provides an emergent event public opinion evolution analysis method based on emotion analysis and theme characteristics.
The invention relates to short text sentiment value calculation and Dynamic Topic Model (DTM) technology, and discloses a novel sentiment analysis-based method for analyzing the evolution of the network public sentiment of an emergency.
The invention aims to solve the technical problems provided by the background technology, adopts the technical scheme that the method for analyzing the public sentiment evolution of the emergency based on the emotional analysis and the theme characteristics comprises the following steps:
1) performing word segmentation processing on the data text: according to the characteristics of Chinese word segmentation, a word segmentation method based on character string matching is adopted to segment the data text;
2) using synonym forest corpus to expand the words after word segmentation, including characteristic words, negative words and degree adverbs, on weight and characteristics and carrying out emotion value calculation;
3) calculating the absolute value of the emotion value difference between adjacent texts, and performing time slicing on data based on the emotion difference;
4) the data after time slicing is used as the input of a Dynamic Topic Model (DTM), and then the change and the heat of the subject word of the sudden event on a time sequence are obtained;
5) analyzing the reasons for the occurrence of the public sentiment changes of the emergency by using the obtained subject words, and fitting the development trend of the emergency by using the heat of the subject words so as to predict the development trend of the emergency.
The step 1) of the invention is to combine a second-order Markov chain to carry out word segmentation processing on the text document.
Step 2) of the invention is to perform emotion value calculation on the short text of the emergency on the basis of step 1), and the characteristics of the network text need to be considered, wherein the characteristics of the network text are ①, ②, timeliness and ③ ambiguity.
Step 3) of the present invention is to determine the time point of time slicing the data based on step 2), and specifically comprises:
(1) calculating the absolute value of the difference value of the emotion values of two adjacent data;
(2) time points are arranged in a descending order, so that the time points meeting the requirements are found by analogy, and the average value of the time points which are closer is calculated as one time point;
(3) and continuing to add new nodes in the sequencing time points to continue the calculation until the conditions are met.
The invention also comprises the step of verifying the role of emotion analysis in the topic evolution process by contrasting the topic evolution result only considering the text similarity segmentation.
The invention tries to introduce a short text sentiment analysis technology in a natural language processing technology into a network public sentiment evolution analysis method. A microblog public opinion evolution analysis method fusing theme and emotional characteristics is planned to be constructed, a synergetic rule between a microblog public opinion theme and emotion of an emergency is disclosed, and reasons generated by theme change in the evolution process of the emergency are analyzed. And verifying the role of emotion analysis in the topic evolution process by contrasting the topic evolution result only considering the text similarity segmentation.
Advantageous effects
1. The method is used for researching the change of various emotional tendencies along with the time shift in public opinion analysis. Compared with the traditional method for researching the public sentiment development situation by aiming at the change of the observation value of the public sentiment quantity, the method can find the time point of the public sentiment change earlier and better fit the evolution trend of the network public sentiment of the emergency. In addition, compared with the traditional method, the method integrates the theme characteristics, and mining aiming at the theme of the emergency and the corresponding emotion of each theme can help organizations such as government enterprises and the like to quickly know the public emotion trend and proportion change in the event occurrence process, regulate, control and predict the emergency, judge the information demand and focus of the public and further make quick and timely response, judge the emotional tendency and emotion aggregation degree in the emergency in time, and help the organizations to adopt effective measures to dredge in time, so that the emotion polarization phenomenon is avoided.
2. The emotion of netizens is analyzed in a fine granularity mode, the proportion change and the trend of the emotion of the public sentiment in the occurrence process can be mastered, and the method has important practical significance for prediction and regulation of the public sentiment.
3. The method is used for constructing the microblog public opinion evolution analysis method fusing emotion and theme characteristics, disclosing the cooperative rule between the microblog public opinion theme and the emotion of the emergency, and providing a scientific and reasonable decision basis for the management department of the emergency in the aspects of public opinion judgment and risk prediction.
Drawings
FIG. 1 is a schematic flow diagram of the process.
FIG. 2 is a Dynamic Topic Model (DTM) process.
FIG. 3 is a sequential evolution diagram of microblog of Zika virus.
FIG. 4 shows keyword popularity fitting and evolution trend.
FIG. 5 shows distribution of topic terms in each stage of a microblog Zika virus event development cycle.
Detailed Description
To further explain the implementation of the present invention in detail, the present invention is developed by combining the embodiments and the attached drawings.
The specific steps of the invention are developed as follows:
step S0101: the word segmentation processing is carried out on the data text, and the existing Chinese text word segmentation algorithm can be classified into three categories according to the characteristics of Chinese word segmentation: the first category is a word segmentation method based on string matching, the second category is a word segmentation method based on understanding, and the third category is a word segmentation method based on statistics. The invention relates to a word segmentation method based on character string matching, which combines a second-order Markov chain to perform word segmentation processing on a text document.
In the process of text emotion analysis, word segmentation processing of a text is carried out firstly, and a word generation model is used in the process. The word segmentation process used in the present invention uses the following word generation model, as shown in equation (1):
wherein: w Seq ≡ ω1 m＝[ω1,ω2,...ωm]Representing a set of m words, c1 nRepresenting a sentence containing n words, the word-trigram model is represented as follows: p (omega)i|ωi-2,ωi-1) The concrete expression method is shown as formula (2):
step S0102: when P (c)1 n|ω1 m) 1 and P (c)1 n) Equal to W Seq, consider only P (ω)1 m) Further, the method can be simplified into the setting of a second-order Markov chain;
the second order Markov chain settings are shown in equation (3):
step S0201, carrying out sentiment value calculation on short texts of network public sentiments, wherein the characteristics of the network texts need to be considered, the characteristics of ① texts are short and irregular, the timeliness is ② and the ambiguity is ③, the method utilizes corpora such as synonym forest and the like to carry out sentiment value calculation on the participled words on the basis of weight and characteristics by utilizing an expression (1):
S(f)＝neg*deg ree*wf(f) (4)
wherein S (f) represents the sentiment value of the feature f; w is af(f) Is the weight of the feature; neg represents whether the characteristic has a modified negative word, if the characteristic has the modified negative word, neg is equal to-1, otherwise neg is equal to 1; deg re represents the degree adverb of the modified feature, if the feature is not modified by the degree adverb, deg re is 1, otherwise, deg re is assigned as shown in table 1:
TABLE 1 weight correspondence table of partial degree adverbs and negatives
For a piece of text, its emotion value can be represented by equation (5):
wherein: s(s) represents an emotion value of the text s, F represents a feature set included in the text s, F represents a feature of the text, s (F) is an emotion value of the feature F calculated by the formula (1), and count (F) represents a feature number of the text s.
Step S0301: time-slicing based on emotion values considers emotion values as features of the slice. Firstly, calculating the absolute value of the difference value of the emotion values of two adjacent data, then arranging time points in a descending order, finding out the time point meeting the requirement by analogy, and calculating the average value of the time points which are closer to each other to be used as a time point; and continuing to add new nodes in the sorting time points to continue the calculation until the conditions are met.
The specific flow is as follows:
step S0401: dynamic Topic Model (DTM), which allows the detection and tracking of current and periodic interests and changes in topics and emotions. Each document has its corresponding topic distribution, and a word distribution corresponding to each topic. The change of the theme in the evolution process of the corresponding emergency can be obtained by analyzing the theme distribution corresponding to the document, and the reason of the public sentiment change in the evolution process of the emergency can be obtained by analyzing the word distribution corresponding to the theme. The development trend of the emergency can be fitted by utilizing the heat of the subject word, and the development trend of the emergency can be predicted. The Dynamic Topic Model (DTM) performs the following:
(1) assuming a total of m documents, a total of k topics are involved
(2) Each document has its own topic distribution subject to a plurality of distributions whose parameters are considered to follow its conjugate distribution Dirichlet distribution whose parameter is α
(3) Each topic has its own word distribution, which obeys a multinomial distribution whose parameters are considered to obey a Dirichlet distribution whose parameters are β
(4) For the nth word in a certain document, firstly sampling a theme from the theme distribution of the article, then sampling a word in the word distribution corresponding to the theme until the processes are completely finished for m documents
Step S0402: using the joint probability distribution of topics and words, as shown in equation (6), and finding the maximum likelihood estimate thereof, as shown in equation (7):
by integrating sum of thetaEliminating to obtain: p (w, z | α, β), and P (z | w, α, β) is solved to P (w, z | α, β). Where z is the hidden variable (i.e., topic) corresponding to each word in the corpus, w is the document corresponding to each topic, θ is the topic distribution of each document,is the word distribution of each topic, randomly initializes z, theta and
may be derived by likelihood estimation.
Step S0403: because the samples of the probability distribution P (z | w, α, β) are difficult to generate, P (z | w, α, β) samples are generated using Gibbs sampling. The training process of the DTM Gibbs sampling algorithm is as follows:
① selecting proper topic number k, selecting proper hyper-parameter vector alpha, beta, ② corresponding to each word of each document in the corpus, randomly assigning a topic number z, ③ rescanning the corpus, updating the topic number of each word by using Gibbs sampling formula and updating the number of the word in the corpus, ④ repeating the Gibbs sampling based on rotation of coordinate axes in step ② until the Gibbs sampling converges, ⑤ counting the topics of each word in the corpus to obtain the topic distribution theta of the document, counting the distribution of each topic word in the corpus to obtain the topic and word division of the DTMCloth
The prediction process of the DTM Gibbs sampling algorithm is as follows:
① corresponding to each word of the current document, randomly assigning a topic number z. ② rescanning the current document, and for each word, updating its topic number by using Gibbs sampling formula. ③ repeating the Gibbs sampling based on rotation of coordinate axes in step ② until the Gibbs sampling converges. ④ counts the topics of each word in the document to obtain the document topic distribution.
Step S0404: in the method for solving DTM by the Gibbs sampling algorithm, alpha and beta are known prior inputs, and the aim is to obtain each prior input
wknThe probability distribution of the corresponding whole z, w, i.e. the distribution of the document subject and the distribution of the subject words. Due to the adoption of the Gibbs sampling method, for the required target distribution, conditional probability distribution of each characteristic dimension of the corresponding distribution needs to be obtained, and the conditional probability distribution is shown as a formula (8).
Step S0405: the joint distribution P (w, z) of w and z is obtained, and a certain word w can be obtainediCorresponding topic feature ziIs determined according to the conditional probability distribution P (zi ═ k | w, z). With the conditional probability distribution P (zi ═ k | w, z), we can perform Gibbs sampling, and finally obtain the topic of the i-th word after the Gibbs sampling converges, where the conditional probability formula of the Gibbs sampling of each word corresponding to the topic is shown in formula (9).
Wherein
Indicating the number of kth topics in the d-th document,
indicates the number of the t-th words in the k-th subject, alpha and beta respectively indicate the subject distribution theta and the word distribution
Is determined.
And (3) the topics of all words can be sampled by using the formula (9), and the sampling topics of all words can be obtained after Gibbs sampling is converged. By using the corresponding relation between the words and the topics obtained by sampling, the distribution theta of the word topic of each document and the distribution of all words in each topic can be obtained
Step S0501: and verifying whether the method based on the sentiment value fragmentation can fit the development trend of public sentiment evolution by using the heat of the subject words, and finding out the time point of the public sentiment change. The reason of the public opinion change in the emergency can be analyzed and obtained by utilizing the theme distribution of the document and the word distribution corresponding to the theme.
Example (b):
FIG. 2 is a Dynamic Topic Model (DTM) process where z is the hidden variable (i.e., topic) for each word in a corpus, θ is the topic distribution for each document,
is the word distribution of each topic, randomly initializes z, theta and
may be derived by likelihood estimation.
Take the example of a microblog emergency "Zika virus". Firstly, a public opinion evolution trend curve of the microblog Zika virus events on a time sequence is counted, as shown in fig. 3. It can be seen from the figure that the Zika virus event occurred from 2016 at 9/30, and the number of subject microblogs continued to increase over time, with a sharp trend at 10/8 and 10/13 and peaked at 10/18.
By adopting an online public opinion evolution method based on emotion analysis, time slicing is performed on Zika virus data respectively based on emotion values and text similarity, and theme word popularity analysis is performed at the time points of the time slicing, as shown in FIG. 4.
As can be seen from fig. 4, compared with time slicing of data based on text similarity, time slicing of data based on emotion values can better fit the actual development and evolution trend of the emergency public sentiment. And the time point of public sentiment change can be found earlier compared with the text similarity method.
As can be seen from FIG. 4, in the process of the evolution of the microblog Zika virus event, the trend of the Zika virus event is influenced along with the change of the theme. From the distribution of subject terms at each stage, it can be seen that the reason for the initiation of this event is the appearance of many "cases of Zika virus" in Europe. The reason for the second growth is "Olympic Congress in the Seika virus disaster area Brazil.
The invention tries to introduce a short text sentiment analysis technology in a natural language processing technology into a network public sentiment evolution analysis method. A microblog public opinion evolution analysis method fusing theme and emotional characteristics is planned to be constructed, a collaborative rule between a microblog public opinion theme and emotion of an emergency is disclosed, the theme distribution of a document and the word distribution corresponding to the theme analyze the change of the emergency public opinion to generate a cause, and the development trend of the emergency is fitted by using the heat of the theme words, so that the development trend of the emergency is predicted. In addition, the topic evolution result of the text similarity segmentation is only considered in a contrast mode, and the role of emotion analysis in the topic evolution process is verified.
Claims (5)
1. The method for analyzing the public sentiment evolution of the emergency based on the emotional analysis and the theme characteristics is characterized by comprising the following steps of:
1) performing word segmentation processing on the data text: according to the characteristics of Chinese word segmentation, a word segmentation method based on character string matching is adopted to segment the data text;
2) utilizing synonym forest corpus to expand the words after word segmentation, including characteristic words, negative words and degree adverbs, on the weight and the characteristics and carry out emotion value calculation;
3) calculating the absolute value of the emotion value difference between adjacent texts, and performing time slicing on data based on the emotion difference;
4) the data after time slicing is used as the input of a Dynamic Topic Model (DTM), so as to obtain the change and the heat of the subject words of the emergency on a time sequence;
5) analyzing the reasons for the occurrence of the public sentiment changes of the emergency by using the obtained subject words, and fitting the development trend of the emergency by using the heat of the subject words so as to predict the development trend of the emergency.
2. The method for analyzing the public opinion evolution of emergency based on emotional analysis and topic features as claimed in claim 1, wherein the step 1) is to perform word segmentation on the text document by combining a second-order Markov chain.
3. The method for analyzing the public sentiment evolution of emergency events based on emotional analysis and topic features as claimed in claim 1, wherein the step 2) is to perform sentiment value calculation on the short text of the emergency event based on the step 1), and the characteristics of web text are considered, wherein the characteristics of the web text are ① text is short and irregular, ② timeliness and ③ ambiguity.
4. The method for analyzing the evolution of the public sentiment of the emergency based on the emotional analysis and the topic characteristics as claimed in claim 1, wherein the step 3) is to determine the time point of time slicing for the data based on the step 2), and specifically comprises:
(1) calculating the absolute value of the difference value of the emotion values of two adjacent data;
(2) time points are arranged in a descending order, so that the time points meeting the requirements are found by analogy, and the average value of the time points which are closer is calculated as one time point;
(3) and continuing to add new nodes in the sequencing time points to continue the calculation until the conditions are met.
5. The method for analyzing the public sentiment evolution of emergency events based on emotional analysis and topic features as claimed in claim 1, further comprising verifying the role of emotional analysis in the topic evolution process by comparing the topic evolution results considering only the text similarity segment.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810718522.3A CN110728140A (en) | 2018-06-30 | 2018-06-30 | Emotion analysis and theme feature-based emergent event public opinion evolution analysis method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810718522.3A CN110728140A (en) | 2018-06-30 | 2018-06-30 | Emotion analysis and theme feature-based emergent event public opinion evolution analysis method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110728140A true CN110728140A (en) | 2020-01-24 |
Family
ID=69216793
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810718522.3A Pending CN110728140A (en) | 2018-06-30 | 2018-06-30 | Emotion analysis and theme feature-based emergent event public opinion evolution analysis method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110728140A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111309864A (en) * | 2020-02-11 | 2020-06-19 | 安徽理工大学 | User group emotional tendency migration dynamic analysis method for microblog hot topics |
| CN112182187A (en) * | 2020-09-30 | 2021-01-05 | 天津大学 | Method for extracting important time segments in short text of social media |
| CN114742281A (en) * | 2022-03-21 | 2022-07-12 | 西北工业大学 | Public event network public opinion popularity prediction method based on grey model |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107315778A (en) * | 2017-05-31 | 2017-11-03 | 温州市鹿城区中津先进科技研究院 | A kind of natural language the analysis of public opinion method based on big data sentiment analysis |
-
2018
- 2018-06-30 CN CN201810718522.3A patent/CN110728140A/en active Pending
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107315778A (en) * | 2017-05-31 | 2017-11-03 | 温州市鹿城区中津先进科技研究院 | A kind of natural language the analysis of public opinion method based on big data sentiment analysis |
Non-Patent Citations (2)
| Title |
|---|
| 何天翔: "基于情感词网的短文本情感分类方法研究" * |
| 何天翔等: "一种基于情感分析的网络舆情演化分析方法" * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111309864A (en) * | 2020-02-11 | 2020-06-19 | 安徽理工大学 | User group emotional tendency migration dynamic analysis method for microblog hot topics |
| CN111309864B (en) * | 2020-02-11 | 2022-08-26 | 安徽理工大学 | User group emotional tendency migration dynamic analysis method for microblog hot topics |
| CN112182187A (en) * | 2020-09-30 | 2021-01-05 | 天津大学 | Method for extracting important time segments in short text of social media |
| CN112182187B (en) * | 2020-09-30 | 2022-09-02 | 天津大学 | Method for extracting important time segments in short text of social media |
| CN114742281A (en) * | 2022-03-21 | 2022-07-12 | 西北工业大学 | Public event network public opinion popularity prediction method based on grey model |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Saad et al. | Twitter sentiment analysis based on ordinal regression | |
| Ni et al. | Sentiment Analysis based on GloVe and LSTM-GRU | |
| Zhang et al. | Comparison of text sentiment analysis based on machine learning | |
| Li et al. | News text classification model based on topic model | |
| CN106599065B (en) | Food safety network public opinion early warning system based on Storm distributed framework | |
| CN112307153B (en) | Automatic construction method and device of industrial knowledge base and storage medium | |
| CN110728140A (en) | Emotion analysis and theme feature-based emergent event public opinion evolution analysis method | |
| CN113962293A (en) | LightGBM classification and representation learning-based name disambiguation method and system | |
| Wan | Sentiment analysis of Weibo comments based on deep neural network | |
| Chang et al. | A METHOD OF FINE-GRAINED SHORT TEXT SENTIMENT ANALYSIS BASED ON MACHINE LEARNING. | |
| Hannah et al. | Automatic extractive text summarization based on fuzzy logic: a sentence oriented approach | |
| Xiao et al. | Patent text classification based on naive Bayesian method | |
| Jawad et al. | Combination Of Convolution Neural Networks And Deep Neural Networks For Fake News Detection | |
| Zhang et al. | An effective convolutional neural network model for Chinese sentiment analysis | |
| Li | Forecast and simulation of the public opinion on the public policy based on the markov model | |
| Ding et al. | Sentiment analysis of chinese micro-blog based on classification and rich features | |
| Zhang et al. | Query classification using convolutional neural networks | |
| CN106202116A (en) | A kind of file classification method based on rough set and KNN and system | |
| Li et al. | A new vector representation of short texts for classification | |
| Wang | Research on the art value and application of art creation based on the emotion analysis of art | |
| Wang et al. | A Method of Hot Topic Detection in Blogs Using N-gram Model. | |
| Li | Deep Learning-Based Natural Language Processing Methods for Sentiment Analysis in Social Networks | |
| Imran et al. | Twitter Sentimental Analysis using Machine Learning Approaches for SemeVal Dataset | |
| Kaur et al. | Sentiment analysis of english tweets using data mining | |
| Zuo et al. | Data mining strategies and techniques of internet education public sentiment monitoring and analysis system |
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: 20200124 |