/QWIM_MLHub

QWIM Project: Applications of machine learning for empirical asset pricing and risk premia forecasting

Primary LanguageJupyter Notebook

Machine learning for index price prediction and portfolio weight optimization

For NYU industrial-hosted project - BofA QWIM(Quantitative Wealth and Investment Management)

Team Information:

Team Name: MLHub
Team Member: Xin Gu xg848@nyu.edu, Dingtian Zhu dz1388@nyu.edu

Description

Here we want to use machine learning methods, mainly to build a LSTM-RNN model, to realize an automated prediction for index price using historical daily price data. The model would mainly help investors in asset allocation and portfolio construc- tion. Stock-level predictive characteristics and macro-economical factors could be added on the base of our model to build more complicated ones for stock price prediction or to improve performance.

Dataset used

We load daily price data of 49 indexes from Bloomberg terminal and cleaned data is attached as "dt_2001_clean.csv". Of course you can also scrape data from online sources like Yahoo Finance.