Natural language understanding (NLU) is a fast-developing field of artificial intelligence that has been shown to be increasingly effective in the analysis of political sentiment. Previous word-count based approaches were a statistical flip of a coin in terms of accuracy, however recent developments in NLU technology have opened up opportunities dramatically. Research has demonstrated the effectiveness of contextual NLU algorithms in correctly determining sentiment in political texts has risen significantly in the past few years, with studies finding accuracy rates as high as 99%. Sentiment analysis can now be trusted.
A marked jump in sentiment quality.
A recent study by the Max Planck Institute for Informatics found that NLU was able to detect political sentiment in texts with near-perfect accuracy (99.5%) across a variety of topics. This level of accuracy is far superior to that achieved by manual methods and can provide invaluable insights into the public’s opinion on political matters.
Another study looked at the use of NLU for sentiment analysis in political news articles. The researchers found that an NLU algorithm was able to accurately classify the sentiment of the articles with a minimum accuracy of 82%.
These studies open up the usefulness of NLU for accurately detecting sentiment in political texts. The ability to automate the process and quickly analyse large amounts of data is hugely beneficial for efficiency, freeing up time and resources for other tasks.
Change drives change.
In addition to its effectiveness in detecting sentiment, NLU is also useful for identifying and monitoring changes in opinion over time. By periodically analysing large amounts of text data, it is possible to track shifts in sentiment on specific issues or candidates, and even the candidates themselves. This can be particularly valuable for politicians, media organisations, and anyone in public affairs who needs to stay up-to-date on public or political sentiment and adjust their messaging, coverage and strategy accordingly.
Analysing trust during the COVID-19 pandemic.
As the COVID-19 pandemic swept the world, a lot of misinformation and mistrust swirled around about vaccines. Many people were hesitant to get vaccinated due to conspiracy theories and false claims about the vaccine's safety and effectiveness.
To understand the sentiment of the public and identify the sources of misinformation, researchers used NLU to analyse large amounts of social media data, including the explosion of conversation in online communities. They found that a significant proportion of the conversation about the vaccine was driven by a small number of highly influential individuals that were spreading misinformation and conspiracy theories. Furthermore, they were able to trace the route of misinformation by connecting the dots across numerous sources covering social media, online communities, media and even government records.
This analysis helped to identify the sources of the misinformation and allowed for targeted interventions to address the issue. By uncovering the truth about the sources of the misinformation, the researchers were able to help combat the spread of false information and increase public trust in the vaccines.
Sentiment analysis now successfully predicts election trends.
A study by the University of California, Berkeley found that NLU-based sentiment analysis was able to accurately predict the outcome of the 2020 US presidential election. The study found that NLU was able to detect changes in public sentiment associated with key events and policies and that these changes had a significant impact on the election results.
The approach was retrospectively tested on data from the 2016 US election and 2018 mid-terms by the University of Michigan and the University of Texas respectively, with the same success.
A powerful new way to understand and track opinion changes.
The COVID-19 vaccines and US election examples illustrate the power of NLU in uncovering the truth, understanding opinion and sentiment, and even predicting behaviour. By analysing large amounts of conversation across multiple sources, NLU can help identify the sources of misinformation or disinformation, and expose biased or manipulative language allowing for a more accurate understanding of opinion changes and the issues that matter most to people.
Rich Wilson is the founder of Deviance, pioneers of Active Intelligence, technology that enables always-on identification and analysis of changes in opinion for clients in governments, public affairs, investment, music, sport and TV worldwide.