AI predicts the U.S. election this time it abandoned Trump

Biden is six votes out! Under U.S. election rules, the first candidate to get 270 votes will win the presidency of the United States! On the second day of Election Week, Biden led by 264 votes. So far, he has successfully won Michigan (16 electoral votes), Wisconsin (10 electoral votes) two swing states, and if he goes on to win Nevada by six more votes, Biden will directly ask the new president of the United States.

AI predicts the U.S. election this time it abandoned Trump

Biden’s lead in Nevada has risen from 7,647 to more than 11,000, according to a new CNN report.

By contrast, Mr. Trump lost again, dropping his lead in Georgia to 10,000 votes. Hopes of “turning the page” in three key states, Pennsylvania, North Carolina and Georgia, now look increasingly slim.

Mr. Trump, who has repeatedly lost his seat, couldn’t sit still, even tweeting last night that he would stop counting votes.

AI predicts the U.S. election this time it abandoned Trump

But whatever Mr. Trump may say, Mr. Biden’s tendency to win the presidency of the United States seems irreversible.

In fact, there was no suspense in this US election. Even before the election began, public opinion began to pour into Biden, especially on social platforms such as Twitter.

Recently, Dr. Hernan Makse, a statistical physicist at City University of New York, also used AI and big data technology to analyze popular opinion on social platforms to predict the winner of this election.

And the result is as expected.

AI: Biden asks the president of the United States

Democratic candidate Joe Biden will defeat current U.S. President Donald Trump to become the next president.

This is the conclusion of the AI model developed by Makse.

AI predicts the U.S. election this time it abandoned Trump

“We collected 1 billion tweets and a lot of comments on social platforms like Twitter, and then we entered that data into the AI model to come up with this prediction,” Makse said in a recent media interview. “

Makse is understood to have started the AI study in the 2016 US election and the 2019 Argentine election. The current projections for the 2020 U.S. election are based on 200 million tweets collected from the last U.S. election and 800 million tweets collected this time.

According to Makse, considering that every presidential election, a large number of people express their views on social platforms, they start using this information to train machine learning models to predict the winner in advance.

Specifically, they trained a neural network that uses AI models to predict the opinions of everyone on social platforms. With this information, big data analytics can be used to predict the final outcome of the election.

However, it should be emphasized that this ai combined with big data forecasting methods also has obvious limitations. Makse says that “the model is not yet fully prepared to accurately predict the outcome of the election” and explains two factors:

First, there is sampling bias. The forecast comes mainly from Twitter. But this is objective, because even if you survey 20 million people across the United States, the sample is always biased. In addition, if the content of the tweet is not semantically analyzed, but only the number of statistics, will also significantly affect the final accuracy.

Second, the voter turnout is biased. It is not possible to know exactly how many people actually voted.

These two variables are important for the accuracy of AI predictions. ‘Based on these two points, we are currently using data from previous and current elections to try to calibrate the AI model, ‘ Mr. Makse said. At the same time, he stressed that

Although Biden is the favorite in this election, AI is not yet ready to accurately predict the outcome, but according to the current AI model, Biden is a clear winner.

Social networks influence the U.S. election

The research team that uses AI to predict election results is much more than that.

However, almost all teams have focused their data collection channels on social media platforms, particularly Twitter.

AI predicts the U.S. election this time it abandoned Trump

The AI prediction system, created by scientists at the University of Granada, is understood to have used Twitter data to verify the winner of the 2016 US election.

In other words, AI verified the fact that Trump defeated Hillary Clinton to win the U.S. election by tweeting.

According to the researchers, they collected a large number of political topics and posts on Twitter involving Trump and Hillary Clinton, and then used the data for AI analysis. It turns out that American society is more angry with Mrs Clinton than Mr Trump when it come to emotional analysis. In addition, Mr. Trump has a better connection to “trust” sentiment. This means that the majority of Twitter users who tweet trust and support Trump, and that’s the result of the final election.

AI predicts the U.S. election this time it abandoned Trump

If Biden wins the election, as AI predicted, does it reflect in part the likely impact of social networks such as Twitter on the U.S. election?

In response, Professor Schwarz gave a positive answer, saying,

While our study does not demonstrate the possible impact of social media platforms other than Twitter, such as Facebook, or the potential role of international political interference or misleading information, we do find that social media can influence election results.

In fact, since the 2016 election, the U.S. has been buzzing about the social platform Twitter helped Republican candidate Donald Trump win the presidency of the United States. In a recent presidential interview, Mr. Trump himself said, “I wouldn’t be here without social media.”

Recently, Professor Schwarz’s University of Bocconi University, in partnership with Thomas Fujiwara and Princeton University, conducted a survey on the relationship between social networking and the US election.

They found that, in addition to reflecting the election results to some extent, Twitter could put Donald Trump at a disadvantage.

If the number of Twitter users in a county doubles, Mr. Trump’s share of the vote would drop by about 2 percentage points, Professor Schwarz said.

The conclusion comes after data analysis found that most Twitter users are young, well-educated liberals living in urban areas, while Trump supporters, mostly rural or college-educated, are among the least likely to use social media.

Discussion: Is Biden the best candidate?

So far, Biden has received more than 71 million votes, arguably a record high.

But are Biden really the best candidates for president of the United States for supporters?

AI predicts the U.S. election this time it abandoned Trump

Recently, the editor-in-chief found that there is a common saying on social networks that the 2020 U.S. election is in fact a one-man show for Trump. Most people think that voting for Trump means trusting or agreeing with Trump and his policies.

A significant portion of the vote for Biden was not due to trust and approval of Biden and his policies, but to distrust, dislike and even resentment of Trump.

What do you think of that? Welcome to the comments area?