An Optimized Deep Learning Model for Emotion Classification in Tweets

Natural Language Processing

Sentiment analysis of tweets has become increasingly important due to the wide range of emotions expressed, from national leaders to the average person. This paper proposes a hybrid LSTM-CNN model to improve accuracy and efficiency by comparing various deep learning and machine learning techniques.

This research was conducted during my CS undergraduate minors in Data Science under the capable guidance of Ps Rana at Thapar Institute of Technology.

Research Methodology

• Model Testing: Rigorous testing of various models to classify tweet sentiments.
• Hybrid Approach: Leveraging LSTM’s ability to capture long-term dependencies and CNN’s proficiency in feature extraction, resulting in enhanced precision.
• Data Preprocessing: Initial steps included removing stop words and extraneous data, significantly boosting time efficiency and accuracy compared to traditional methods.

Key Findings

The hybrid model demonstrated superior performance metrics, showcasing its potential for robust sentiment analysis in social media contexts.

Recommended citation: Yash Singh Pathania,(2021). "An Optimized Deep Learning Model for Emotion Classification in Tweets
Download Paper