https://github.com/akashprak/stockprice-timeseries
This repository contains the python code for forecasting stock prices using various Time Series models like ARIMA, SARIMAX.
https://github.com/akashprak/stockprice-timeseries
arima exploratory-data-analysis machine-learning sarimax statsmodels timeseries
Last synced: about 1 month ago
JSON representation
This repository contains the python code for forecasting stock prices using various Time Series models like ARIMA, SARIMAX.
- Host: GitHub
- URL: https://github.com/akashprak/stockprice-timeseries
- Owner: akashprak
- Created: 2025-02-23T17:36:54.000Z (8 months ago)
- Default Branch: main
- Last Pushed: 2025-03-22T12:54:31.000Z (7 months ago)
- Last Synced: 2025-04-12T10:56:10.247Z (6 months ago)
- Topics: arima, exploratory-data-analysis, machine-learning, sarimax, statsmodels, timeseries
- Language: Jupyter Notebook
- Homepage:
- Size: 1.75 MB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# StockPrice-TimeSeries
Time series analysis is used to help model time based data and forecast future prices and trends.
This Project builds a machine learning model which uses historical stock data to predict future prices using various machine learning techniques.
## Dataset
__Source__: Historical stock price and volume of FACT Ltd.
## Methodology
- Data Preprocessing
- EDA
- Stationarity Checks
- ACF and PACF plots
- Differencing
- Model Training - ARIMA
- Model Training - SARIMAX
- Model evaluation
Performance Evaluation is done using RMSE, AIC and BIC.
- Forecasting
### ARIMA
An ARIMA model is trained on the Closing price of the stock data and the best parameters for the model is selected. A rolling forecast is done against the test data and the model with the lowest RMSE is selected.
### SARIMAX
The SARIMAX model is trained on the Close price with the Volume data as the exogeneous variable.
### PROPHET
Prophet is a time-series forecasting tool by Meta which automatically handles trends, seasons, missing values and is robust to outliers. It is a powerful tool for time series forecasting especially when there is complex patterns involved.
PROPHET is trained on the Close price and forecasting is done along the test data.
## Dependencies
- `pandas`
- `numpy`
- `seaborn`
- `matplotlib`
- `statsmodels`
- `scikit-learn`