Grade 11 Student Develops AI Model To Predict Election Results Using Sentiment Polarity Scores
Prajwal Reddy, a class 11th student at Greenwood High International School, has developed an Artificial Intelligence (AI) model that can predict election results using sentiment polarity scores. The model focuses on the engagement ratings of tweets from various Karnataka politicians, assigning a score based on the tweet's content, language, likes, replies, and retweets. Prajwal claims that his model can be used for any state or national elections and that it is a step ahead of previous research since it also considers other factors such as retweets, likes, and comments that measure the engagement of a politician's tweet.
According to Prajwal, datasets are crucial for analysing sentiment polarity, and his AI model has an accuracy rate of 72.7%. He believes that the results can be extended to predict elections in other states and to understand the impact of positive and negative sentiments on a candidate's chances of winning. However, he acknowledges that bots can affect the model's accuracy by decreasing the quality of engagement ratings.
Prajwal's methodology involves conducting sentiment analysis and polarity on each tweet for each candidate. If the tweet is in a local language, it is translated into English to conduct sentiment analysis successfully. Prajwal believes that translating the tweet body does not reduce the conveyed information, as only the sentiment and not the text's pure meaning needs to be conveyed.
In conclusion, Prajwal believes that his AI model has promising future applications and can be used to predict which politician will win each constituency in the upcoming 2024 general elections. However, he notes that data collection is a challenge that needs to be addressed for the model to be successful.