Vaishali Ingle

Unstructured data makes up approximately 85% of all generated data. Although it lacks a predefined format, it contains valuable insights that businesses can utilize. In the stock market, unstructured data—especially from online news sources—can significantly influence price movements. Techniques such as text mining and visualizations like word clouds are used to uncover correlations between news content and market trends. Traditional machine learning models usually achieve stock price prediction accuracies of50% to 65%. However, SPruH model, which combines Hidden Markov Models with Term Frequency-Inverse Document Frequency (TF-IDF), improves accuracy to 70%. Further advancements through Ensemble Deep Learning Frameworks (EDLF-DP) raise prediction accuracy to around 85%.