Grade 11 Student Develops AI Model To Predict Election Results Using Sentiment Polarity Scores
[ad_1]
Prajwal Reddy in his research paper mentions that his analysis is one step ahead of previous research conducted in the way
Prajwal Reddy claims his AI model has a 72.7% accuracy rate. He mentions that the results by AI model can be expanded to help predict elections for other states
A class 11th student Prajwal Reddy developed an Artificial Intelligence model to predict election results. The AI model developed by Prajwal uses sentiment polarity score. Studying at Greenwood High International School, Prajwal Reddy used his AI model to analyse tweets from various Karnataka politicians. The model focuses on engagement ratings of the tweets tweeted by politicians. The model assign a score on the basis of the tweet’s content, language, number of likes, replies, and retweets.
As stated by Prajwal Reddy, his AI model can be adopted for any state or national elections results. Prajwal Reddy in his research paper mentions that his analysis is one step ahead of previous research conducted in the way. Prajwal states that his model goes ahead as it also looks at other factors, including the number of retweets, likes, and comments a tweet has garner. These factors measures the engagement of the tweet tweeted by a politician. While speaking with News18 Exclusively, Prajwal Reddy stated datasets are important to analyse the sentiment polarity.
In his research paper, Prajwal Reddy claims that his AI model has a 72.7% accuracy rate. He mentions that the results by AI model can be expanded to help predict elections for other states, and could potentially help understand the effect of positive and negative sentiments on the winnability of a political candidate. The more the score the more chance of winnability. When asked on how bots can affect the result of his AI empowered model, he said bots will decrease the quality of the engagement rating overall.
Sharing methodology of his AI model, Prajwal Reddy mentions that, sentiment analysis and polarity is conducted on each tweet for each candidate. The text is translated into English if it is in local language to successfully conduct a sentiment analysis. Prajwal Reddy claims that translation of the tweet body does not reduce the information conveyed as only the sentiment, and not the pure meaning of the text, needs to be conveyed.
While concluding, Prajwal mentioned that his AI model proves promising for future applications. Thus tweet engagement can be seen to correlate to a higher chance of winning. He states that his analysis can be applied to tweets from the upcoming 2024 general elections to help predict which politician will win from each constituency. To make the model successful, a better data collection is the need and a challenge for his model.
Read all the Latest Education News here
Comments
Post a Comment