HIGH VOLATILITY STOCK PRICE PREDICTION USING NEURAL NETWORKS BASED ON HISTORICAL STOCK PRICE DATA AND SENTIMENT ANALYSIS
Stock price prediction is one of the most challenging problems in the realm of data science. There are numerous factors responsible for the price changes in stock market. To address this issue, there are two well-known methods which adopted by many people, namely technical and fundamental analysis....
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Project |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/47743 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | Stock price prediction is one of the most challenging problems in the realm of data science. There are numerous factors responsible for the price changes in stock market. To address this issue, there are two well-known methods which adopted by many people, namely technical and fundamental analysis. Technical analysis exploit the mathematical indicators extracted from the historical stock data without take into account the importance of public sentiment towards a particular company. Fundamental analysis takes into account those ignored factors in technical analysis, but this is a qualitative analysis. In this final task, writer exploited several types of neural networks as the prediction model, and using both Boeing historical stock price and text data from Twitter and StockTwits social media during 2019 as the input to the model. The text data is exploited to extract sentiment score which will be used as the additional predictor variable to the prediction model. Several types of neural networks used in this work including multi layer perceptron, convolutional neural network, and long short-term memory network. |
|---|