39th Annual Student Research Conference
Stock Price Prediction Using Time Series and News Sentiment Analysis
Anup Rijal
Prof. Nazmul Shahadat, Faculty Mentor
This project explores whether combination of stock price history with news sentiment can improve stock market prediction. A deep learning model was constructed and trained on one year of data of eight large companies such as Microsoft, Apple, Tesla and NVIDIA. The model takes into account previous price trends as well as sentiment score of news. Predictions were evaluated comparing predicted prices to actual prices using standard error metrics. The application of artificial intelligence to real world financial decision making is shown by using machine learning, finance and natural language processing techniques.
Keywords: Deep Learning, Sentiment Analysis, Natural Language Processing
Topic(s):Computer Science
Statistics
Presentation Type: Oral Presentation
Session: -
Location: MG 2007
Time: