mcf-long-short / statistics-stocks-forecasting

Empirical analysis with financial data (MSFT stock returns) in R, with the goal to produce useful forecasts using univariate, multivariate time series models and volatility models.

Stock returns forecasting

Docs: https://mcf-long-short.github.io/statistics-stocks-forecasting/

Introduction

The goal of the project is to perform all steps/elements of the empirical analysis with financial/economic data, with the goal to produce useful forecasts. The project has four essential phases and several steps within each phase:

1. Data collection and analysis:
2. Building univariate time series model for forecast
3. Building multivariate model for forecast
4. Building volatility model

Stocks which returns we're forecasting is MSFT. Explanatory variables (features) we're using are: S&P500, Nasdaq and copetitors like AAPL, GOOG, IBM and 3M.

This repository represents group project work for course in Statistics and Financial Data Analysis for advanced degree Masters in Computational Finance, Union University.

Project phases:

Each of the project phases has detailed description of all the steps, implementation details, intuition for modeling, interpretation of data analysis, modeling, evaluation and statistical test that were performed. Here are the links for published R notebooks to RPubs:

• Data collection
• Data analysis
• Building univariate time series model for forecast
• Building multivariate model for forecast
• Building volatility model

Running R markdown notebooks

When you clone the repo you may use renv.

To install all the dependencies, run: renv::init().

And when adding new R packages, you can save them in renv with renv::snapshot().