Sentiment analysis regarding childcare issues using Naive Bayes Algorithm / Alis Farhana Zulkipeli

This study applies the Naive Bayes algorithm for sentiment analysis to assess public perceptions of childcare issues, particularly child abandonment and accidents. With the growing volume of childcare-related discussions on social media, efficient sentiment analysis tools are essential for extracting insights. However, the lack of comprehensive methodologies poses challenges for policymakers, childcare providers, and researchers in understanding public concerns and developing effective interventions. To address this, a dataset of 1,079 tweets from X (formerly Twitter) is analyzed. The data undergoes preprocessing steps such as stop-word and emoji removal, tokenization, and feature extraction using Term Frequency-Inverse Document Frequency (TF-IDF). VADER is used for initial sentiment labeling, and the Naive Bayes classifier categorizes sentiments into positive and negative classes. The motivation behind this project is to leverage sentiment analysis for enhancing childcare policies and public awareness. The project aims to enhance childcare policies and public awareness by leveraging sentiment analysis to bridge the gap between public sentiment and policy decisions. The Naive Bayes model achieves 87% accuracy with high precision, recall, and F1 scores using 10-fold crossvalidation, demonstrating its effectiveness in classifying social media sentiments. Future research could explore advanced techniques like Bidirectional Encoder Representations from Transformers (BERT) or Recurrent Neural Networks (RNNs) to improve classification accuracy and contextual understanding. Expanding the dataset to include multilingual content and incorporating topic modeling techniques would further enhance sentiment analysis in childcare-related discourse.