36 Stock Indices and Commodity Prices Time Series

Published: 19-03-2018| Version 1 | DOI: 10.17632/x744mgjpkv.1
Contributor:
Arash Negahdari Kia

Description

Time series in this dataset are used to create an interaction graph of markets and commodities to be used in machine learning prediction models. We used this dataset in our work and introduced a model named HyS3 and an algorithm named ConKruG.

Files

Steps to reproduce

The steps to reproduce are completely described in our paper. The RAW DATA come from yahoo finance, google finance, Federal Reserve Bank of St. Luis Economic Data, OPEC, U.S. Energy Administration and Tehran Stock Exchange official websites. The data are prepared in a way that all missing values are repeated by their previous data and all are in working days frequency without the Saturdays and Sundays. The preparation was done by pandas module in Python. More detailed information is in our paper. The papers information will be published after the acceptance.