https://github.com/sarvsav/dissertation
Bits dissertation on stock price prediction
https://github.com/sarvsav/dissertation
bits-pilani dissertation dissertation-project machine-learning numerical-analysis stock-price-prediction textual-analysis
Last synced: about 2 months ago
JSON representation
Bits dissertation on stock price prediction
- Host: GitHub
- URL: https://github.com/sarvsav/dissertation
- Owner: sarvsav
- Created: 2022-12-22T08:19:40.000Z (over 2 years ago)
- Default Branch: main
- Last Pushed: 2023-04-03T09:54:51.000Z (about 2 years ago)
- Last Synced: 2025-03-30T11:51:12.873Z (3 months ago)
- Topics: bits-pilani, dissertation, dissertation-project, machine-learning, numerical-analysis, stock-price-prediction, textual-analysis
- Language: Jupyter Notebook
- Homepage:
- Size: 25.2 MB
- Stars: 2
- Watchers: 3
- Forks: 3
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# dissertation
Bits dissertation
The value of stocks always fluctuates dramatically over time, making stock market prediction
an intriguing study subject. Investors do two different forms of study before buying a stock.
The fundamental analysis is one of the first approaches and very common. In order to
determine whether to invest or not, investors consider factors such as the intrinsic worth of the
stocks, the state of the market and economy, the political environment, etc. On the other side,
technical analysis evaluates equities by looking at data produced by market activity, such as
previous prices and volumes. Typically of attempting to determine a security's fundamental
worth, technical analysts instead utilize stock charts to spot patterns and trends that could
predict how a stock will act in the future.
Many methods for predicting stock movements have been developed throughout the years.
Initially, stock trend predictions were made using traditional regression techniques. Non-linear
machine learning methods have also been applied since stock data may be characterized as
non-stationary time series data.
With a forget gate present, the LINEAR, POLY is similar to a long short-term memory (SVM),
however it has fewer parameters than the SVM since it lacks an output gate. The vanishing
gradient issue that arises when using a conventional scaler is addressed with LINEAR, POLY.
The time sequence is erratic and disordered. Most of the forecasting model that uncovers the
complex connection between financial information about an industry and its stock price is
beneficial. The financial news in addition to the existing records concerning the firm is used to
forecast future stock prices.
Semantic and linguistic traits may be extracted using a variety of ways. The following are a
few of them: OpinionFinder, SentiWordNet, Linguistic Inquiry and Word Count (LIWC),
Google Profile of Mood States (GPOMS), R sentiment analysis, and Python NLP package. In
this approach, the sentimental score is also calculated based on news headlines, in addition to
the statistical data for the model to produce more reliable results.
