Discussing the article: "Reimagining Classic Strategies (Part V): Multiple Symbol Analysis on USDZAR"

[MetaQuotes]

Check out the new article: Reimagining Classic Strategies (Part V): Multiple Symbol Analysis on USDZAR.

In this series of articles, we revisit classical strategies to see if we can improve the strategy using AI. In today's article, we will examine a popular strategy of multiple symbol analysis using a basket of correlated securities, we will focus on the exotic USDZAR currency pair.

For us to assess the relationship, we exported all our market data from our MetaTrader 5 terminal using a script written in MQL5. We trained various models using 2 groups of possible inputs for the models:

1. Ordinary OHLC quotes on the USDZAR.
2. A combination of oil and gold prices.

From the data collected, it appears that oil has stronger correlation levels with the UDZAR currency pair than gold.

Since our data were on different scales, we standardized and normalized the data before training. We performed 10-fold cross validation without random shuffling to compare our accuracy across the different sets of inputs.

Author: Gamuchirai Zororo Ndawana

[linfo2]

Awesome , another excellent walk through . Thank you for the workbook , we now have a template to test our own correlations. Much appreciated 

[Gamuchirai Zororo Ndawana]

linfo2 #:
Awesome , another excellent walk through . Thank you for the workbook , we now have a template to test our own correlations. Much appreciated 

The pleasure is mine Neil,  I remember you once told me that you had an idea that involves searching for correlation between the indicators, feel free to share how your findings in that project may help us here, and hopefully we can cook 💯🔥 

[Carl Schreiber]

In the script you aren't closing the file opened to write the date, both in the article and in fetchData.mq5

[Gamuchirai Zororo Ndawana]

Carl Schreiber #:
In the script you aren't closing the file opened to write the date, both in the article and in fetchData.mq5

Thank you for highlighting that, I'll correct myself in future