Unveiling Sentiments: CNN-LSTM Based Social Media Sentiment Analysis

Many people use online social networks to share their opinions on a wide range of topics. In this age of instant global contact, this is especially useful. In many fields, like public opinion studies, brand management, and market research, it’s very important to understand and completely look into these feelings. Combining a CNN and LSTM network is used in this study to propose a method for mood analysis. Feelings of happiness, neutrality, and sadness can be correctly distinguished by the framework. Our method for analyzing sentiment works very well across all types of sentiment, according to the results. A 98% accuracy rate, a 96.22% precision rate, a 97.69% memory rate, and an F1 score of 96.95% are what the model gets when it is used on the Neutral mood group. With regards to happy feelings, the model gets a 99% Accuracy, a 97.96% F1 Score, a 98.06% Recall, and a 97.85% Precision. Findings from the model show a trend for the Negative mood category: 97.74% Accuracy, 97.64% Precision, 97.69% F1 Score, and 98% Recall. Based on all of these results, it’s clear that the suggested approach for mood analysis does a good job of categorizing the feelings people tweet about. Our approach utilizes advanced machine learning methods, including CNN and LSTM networks, to accurately identify subtle emotional reactions and provide valuable insights into human thinking processes and current trends.