NLP: topic LDA, emotion analysis and news text data under the epidemic situation

Time:2022-4-27

Original link:http://tecdat.cn/?p=12310

Original source:Tuoduan data tribe official account

The outbreak of COVID-19 made this year’s Spring Festival different from usual. At the same time, the news recorded the timeline of the development of the epidemic.

Therefore, we analyzed the data on the news content related to the epidemic, the release period, the theme and emotional tendency of the release content, and hope to know more about the epidemic through these data.

Emotional tendency of news on epidemic related topics

By analyzing the theme and emotion of the news related to the epidemic, we can get the keyword and emotion distribution of each theme.

Chart 1

NLP: topic LDA, emotion analysis and news text data under the epidemic situation

The news content of the topic of symptom detection expresses the most positive emotions. Under this topic, the symptoms of patients tested in hospitals are discussed, followed by the news content related to urban services and schools. Topics such as store closure, community isolation and school delay are discussed. The main topic of life also expresses more positive emotions (key words: time and family). The epidemic situation increases the time for families to get along with each other (Fig. 1).

The emotional tendency of news expression changes with time

Considering the time of news release and news related topics, figure 2 shows the results obtained through emotional cross analysis.

Chart 2

NLP: topic LDA, emotion analysis and news text data under the epidemic situation

From the topic ranking, the hottest topics in the news at different times include economy, travel and politics. From the perspective of emotional distribution, the economic theme news in January expressed more negative emotions (for example, the stock market fell due to increasing concern about Coronavirus). In March, with the gradual improvement of the epidemic situation, the heat ranking of urban theme news (such as ensuring store services and production and operation during the epidemic) exceeded the protection theme (key words: mask, hand washing, health, etc.). From January to March, the proportion of positive emotions under each theme is increasing.

News attention to different subject keywords

Considering the attention of different topics, figure 3 shows the distribution of high-frequency keywords.

Chart 3

NLP: topic LDA, emotion analysis and news text data under the epidemic situation

From this, we can see that the most concerned aspects in the news related to the epidemic are health, family, isolation and travel, of which health appears most frequently. Then focus on topics including coronavirus, work during the epidemic and virus detection. Second, the topic of concern includes distinguishing between health and infection symptoms. Other popular keywords of concern include school, business, travel and economy.

All information in this article (including but not limited to analysis, prediction, suggestions, data, charts, etc.) is for reference only, and tecdat will not be liable for any loss arising from all or part of the article or caused by this article.


NLP: topic LDA, emotion analysis and news text data under the epidemic situation

The most popular opinion

1.Little red book user behavior data collection insight: Wedding Grass Planting Guide

2.Machine learning boosts fast fashion and accurate sales forecast

3.Cities on bicycles: insights into sharing bicycle data

4.Using machine learning to identify changing stock market conditions — Application of hidden Markov model (HMM)Application “)

5.Data inventory: new trend of online consumption of household appliances

6.GAM (generalized additive model) is used for power load time series analysis in R language

7.Gene Exploration of Tiger flutter Forum: insight into social user behavior data

8.Grasp the data pulse of taxi driving

9.Smart door lock data chopping strategy